diff --git a/.github/workflows/deltawire-lab.yml b/.github/workflows/deltawire-lab.yml
new file mode 100644
index 000000000..3d1982bcf
--- /dev/null
+++ b/.github/workflows/deltawire-lab.yml
@@ -0,0 +1,24 @@
+name: Verify DeltaWire lab
+
+on:
+  pull_request:
+    paths: ["labs/20-deltawire/**", "go.work", ".github/workflows/deltawire-lab.yml"]
+  push:
+    branches: [main]
+    paths: ["labs/20-deltawire/**", "go.work", ".github/workflows/deltawire-lab.yml"]
+
+permissions:
+  contents: read
+
+jobs:
+  validate:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v7
+      - uses: actions/setup-go@v7
+        with:
+          go-version-file: labs/20-deltawire/go.mod
+      - uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+      - run: bash labs/20-deltawire/scripts/validate.sh
diff --git a/.github/workflows/deltawire-preflight-v4.yml b/.github/workflows/deltawire-preflight-v4.yml
new file mode 100644
index 000000000..76b297723
--- /dev/null
+++ b/.github/workflows/deltawire-preflight-v4.yml
@@ -0,0 +1,24 @@
+name: Verify DeltaWire preflight v4
+
+on:
+  push:
+    branches: [eval/deltawire-preflight-v4]
+
+permissions:
+  contents: read
+
+jobs:
+  validate-v4:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v7
+        with:
+          fetch-depth: 0
+      - uses: actions/setup-go@v7
+        with:
+          go-version-file: labs/20-deltawire/go.mod
+      - uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+      - run: bash labs/20-deltawire/scripts/validate.sh
+      - run: bash labs/20-deltawire/eval/scripts/v4/validate.sh
diff --git a/go.work b/go.work
index 6719b64d1..7c1a66deb 100644
--- a/go.work
+++ b/go.work
@@ -9,4 +9,5 @@ use (
 	./labs/12-product-engineering-loop/product-engineering-loop
 	./labs/15-pitot/pitot
 	./labs/18-settle/settle
+	./labs/20-deltawire
 )
diff --git a/labs/20-deltawire/.gitignore b/labs/20-deltawire/.gitignore
index a2dea5e05..a7963795a 100644
--- a/labs/20-deltawire/.gitignore
+++ b/labs/20-deltawire/.gitignore
@@ -3,3 +3,7 @@
 /testdata/generated/
 /inspect.json
 /inspect.md
+/eval/.generated/
+/eval/tasks/*/environment/.generated/
+/eval/probes/*/environment/.generated/
+/harbor/archive/local-preflight-v0/
diff --git a/labs/20-deltawire/README.md b/labs/20-deltawire/README.md
new file mode 100644
index 000000000..9d29cb69b
--- /dev/null
+++ b/labs/20-deltawire/README.md
@@ -0,0 +1,3 @@
+# DeltaWire
+
+In an exploratory 24-run pilot, the DeltaWire treatment used 24.2% fewer aggregate input-plus-output tokens than control. Both arms received full reward under the pilot’s existing verifiers. Semantic oracles, per-run pairing, and treatment-compliance evidence are being hardened before the confirmatory run.
diff --git a/labs/20-deltawire/THIRD_PARTY_NOTICES.md b/labs/20-deltawire/THIRD_PARTY_NOTICES.md
index ccd236f4e..4ac0eb233 100644
--- a/labs/20-deltawire/THIRD_PARTY_NOTICES.md
+++ b/labs/20-deltawire/THIRD_PARTY_NOTICES.md
@@ -4,8 +4,8 @@ DeltaWire uses the following third-party dependencies:
 
 ## github.com/santhosh-tekuri/jsonschema/v6
 
-- License: Apache License 2.0 (or MIT based on repo)
+- License: Apache License 2.0
 - Version: v6.0.2
 - Purpose: JSON Schema Draft 2020-12 validation
 
-This dependency was chosen because it provides robust Draft 2020-12 support in Go and does not require CGO.
+This dependency was chosen because it provides robust Draft 2020-12 support in Go and does not require CGO.
\ No newline at end of file
diff --git a/labs/20-deltawire/assets/init/INSTRUCTIONS.md b/labs/20-deltawire/assets/init/INSTRUCTIONS.md
deleted file mode 100644
index 91a76cd14..000000000
--- a/labs/20-deltawire/assets/init/INSTRUCTIONS.md
+++ /dev/null
@@ -1,27 +0,0 @@
-# DeltaWire repository instructions
-
-Use DeltaWire for repetitive generated test, fixture, benchmark, or evaluation
-data.
-
-Do not manually generate or edit a managed output when the data can be derived
-from a DeltaWire plan.
-
-Workflow:
-
-1. Create or edit a `.deltawire/plans/*.dw.json` generation plan.
-2. Keep the record schema under `.deltawire/schemas/`.
-3. Run `deltawire validate <plan>`.
-4. Run `deltawire inspect <plan> --format markdown` when preparing a coding plan.
-5. Run `deltawire render <plan>`.
-6. Run `deltawire check <plan>` before claiming completion.
-
-Never use `deltawire render --force` without explicit human approval.
-
-Do not claim model-token savings from byte-amplification measurements.
-
-Generated data is complete only when:
-
-- record-schema validation passes
-- dataset assertions pass
-- the output hash matches repository state
-- `deltawire check` exits successfully
diff --git a/labs/20-deltawire/docs/00-grounding.md b/labs/20-deltawire/docs/00-grounding.md
index 0ad704dbc..dcd3d4087 100644
--- a/labs/20-deltawire/docs/00-grounding.md
+++ b/labs/20-deltawire/docs/00-grounding.md
@@ -2,9 +2,9 @@
 
 | Claim | Evidence | Verification command |
 | --- | --- | --- |
-| Directory | `/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire` | `pwd` |
+| Directory | `labs/20-deltawire` | `pwd` |
 | Is Git Repo | Yes (part of intelligence-flow) | `git rev-parse --show-toplevel` |
-| Git Branch | `lab/19-revector-v0` | `git branch --show-current` |
+| Git Branch | `lab/20-deltawire-agent-eval` | `git branch --show-current` |
 | Git Remote | `origin https://github.com/operatorstack/intelligence-flow.git` | `git remote -v` |
-| Go Version | `go version go1.26.5 darwin/arm64` | `go version` |
-| Go Stable | `go1.26.5` | `go env GOVERSION` |
+| Go Version | `go1.22+` | `go version` |
+| Go Stable | `go1.22+` | `go env GOVERSION` |
\ No newline at end of file
diff --git a/labs/20-deltawire/eval/analysis/analyze_paired.py b/labs/20-deltawire/eval/analysis/analyze_paired.py
index 968757817..7da3523cf 100644
--- a/labs/20-deltawire/eval/analysis/analyze_paired.py
+++ b/labs/20-deltawire/eval/analysis/analyze_paired.py
@@ -1 +1,144 @@
-print('analyze_paired.py stub')
\ No newline at end of file
+import json
+import os
+import sys
+import math
+
+def task_clustered_bootstrap(reductions_by_task, num_samples=1000, seed=42):
+    import random
+    random.seed(seed)
+
+    tasks = list(reductions_by_task.keys())
+    if not tasks:
+        return [0, 0]
+
+    means = []
+    for _ in range(num_samples):
+        # resample tasks with replacement
+        sampled_tasks = random.choices(tasks, k=len(tasks))
+
+        # pool all reductions from the sampled tasks
+        sampled_reductions = []
+        for t in sampled_tasks:
+            sampled_reductions.extend(reductions_by_task[t])
+
+        if sampled_reductions:
+            means.append(sum(sampled_reductions) / len(sampled_reductions))
+
+    if not means:
+        return [0, 0]
+
+    means.sort()
+    lower_idx = int(0.025 * len(means))
+    upper_idx = int(0.975 * len(means))
+
+    return [means[lower_idx], means[upper_idx]]
+
+def main():
+    if len(sys.argv) < 2:
+        print("Usage: python3 analyze_paired.py <extracted_runs.json>")
+        sys.exit(1)
+
+    extracted_path = sys.argv[1]
+
+    if not os.path.exists(extracted_path):
+        print(f"No extracted runs found at {extracted_path}.")
+        return
+
+    with open(extracted_path) as f:
+        runs = json.load(f)
+
+    pairs = {}
+    for r in runs:
+        key = r["pair_id"]
+        if key not in pairs:
+            pairs[key] = {}
+        pairs[key][r["arm"]] = r
+
+    analysis = {
+        "aggregate_reduction": 0,
+        "median_paired_reduction": 0,
+        "geometric_mean_paired_reduction": 0,
+        "task_clustered_95_interval": [0, 0],
+        "pair_wins_losses_ties": {"wins": 0, "losses": 0, "ties": 0},
+        "per_task_family_results": {},
+        "all_assigned_run_result": len(runs),
+        "both_pass_pair_result": 0,
+        "d1_mechanism_compliance_rate": 0
+    }
+
+    reductions = []
+    reductions_by_task = {}
+    d1_compliance = 0
+    d1_total = 0
+
+    sum_b0_tokens = 0
+    sum_d1_tokens = 0
+    log_ratios = []
+
+    for key, pair in pairs.items():
+        if "B0" in pair and "D1" in pair:
+            b0 = pair["B0"]
+            d1 = pair["D1"]
+
+            # total_tokens = input_tokens + output_tokens
+            b0_in = b0.get("input_tokens") or 0
+            b0_out = b0.get("output_tokens") or 0
+            d1_in = d1.get("input_tokens") or 0
+            d1_out = d1.get("output_tokens") or 0
+
+            b0_total = b0_in + b0_out
+            d1_total_toks = d1_in + d1_out
+
+            if b0.get("semantic_oracle_result", {}).get("exact_match") == 1 and d1.get("semantic_oracle_result", {}).get("exact_match") == 1:
+                analysis["both_pass_pair_result"] += 1
+
+                if b0_total > 0:
+                    reduction = (b0_total - d1_total_toks) / b0_total
+                    reductions.append(reduction)
+
+                    task = b0["task"]
+                    if task not in reductions_by_task:
+                        reductions_by_task[task] = []
+                    reductions_by_task[task].append(reduction)
+
+                    sum_b0_tokens += b0_total
+                    sum_d1_tokens += d1_total_toks
+
+                    if d1_total_toks > 0:
+                        log_ratios.append(math.log(d1_total_toks / b0_total))
+
+                    # exact positive, negative, or zero paired deltas
+                    if reduction > 0:
+                        analysis["pair_wins_losses_ties"]["wins"] += 1
+                    elif reduction < 0:
+                        analysis["pair_wins_losses_ties"]["losses"] += 1
+                    else:
+                        analysis["pair_wins_losses_ties"]["ties"] += 1
+
+            if d1.get("deltawire_compliance") == "used_successfully":
+                d1_compliance += 1
+            if d1.get("arm") == "D1":
+                d1_total += 1
+
+    if sum_b0_tokens > 0:
+        analysis["aggregate_reduction"] = 1 - (sum_d1_tokens / sum_b0_tokens)
+
+    if reductions:
+        reductions.sort()
+        mid = len(reductions) // 2
+        analysis["median_paired_reduction"] = (reductions[mid] + reductions[~mid]) / 2.0
+
+    if log_ratios:
+        analysis["geometric_mean_paired_reduction"] = 1 - math.exp(sum(log_ratios) / len(log_ratios))
+
+    if reductions_by_task:
+        analysis["task_clustered_95_interval"] = task_clustered_bootstrap(reductions_by_task)
+
+    if d1_total > 0:
+        analysis["d1_mechanism_compliance_rate"] = d1_compliance / d1_total
+
+    with open("labs/20-deltawire/eval/results/canary-v1/paired_analysis.json", "w") as f:
+        json.dump(analysis, f, indent=2)
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/analysis/extract_runs.py b/labs/20-deltawire/eval/analysis/extract_runs.py
index 79ce617c7..73c2cf402 100644
--- a/labs/20-deltawire/eval/analysis/extract_runs.py
+++ b/labs/20-deltawire/eval/analysis/extract_runs.py
@@ -1 +1,26 @@
-print('extract_runs.py stub')
\ No newline at end of file
+#!/usr/bin/env python3
+"""Extract grounded Harbor metrics; fidelity is supplied only by structured receipts."""
+import json
+import sys
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[4]
+
+
+def main():
+    if len(sys.argv) != 4: raise SystemExit("usage: extract_runs.py <ledger> <receipt-dir> <output>")
+    ledger = json.loads(Path(sys.argv[1]).read_text(encoding="utf-8")); receipt_dir = Path(sys.argv[2]); extracted = []
+    for entry in ledger.get("runs", []):
+        if entry.get("status") != "completed": continue
+        result = json.loads((ROOT / entry["result_path"]).read_text(encoding="utf-8")); agent = result.get("agent_result") or {}
+        receipt_path = receipt_dir / f"{entry['arm']}-treatment-fidelity.json"
+        fidelity = json.loads(receipt_path.read_text(encoding="utf-8"))
+        extracted.append({"pair_id": entry["pair_id"], "task": entry.get("contract_task", entry["task"]), "arm": entry["arm"],
+                          "repetition": entry["repetition"], "trial_id": result.get("id"),
+                          "input_tokens": agent.get("n_input_tokens"), "output_tokens": agent.get("n_output_tokens"), "cache_tokens": agent.get("n_cache_tokens"),
+                          "semantic_oracle_result": {"exact_match": fidelity["semantic_exact_match"]}, "deltawire_compliance": fidelity["classification"],
+                          "result_path": entry["result_path"], "result_sha256": entry["result_sha256"], "transcript_sha256": entry["trajectory_sha256"], "output_artifact_sha256": entry["output_sha256"]})
+    Path(sys.argv[3]).write_text(json.dumps(extracted, indent=2)+"\n", encoding="utf-8")
+
+
+if __name__ == "__main__": main()
diff --git a/labs/20-deltawire/eval/analysis/render_report.py b/labs/20-deltawire/eval/analysis/render_report.py
index 37d69bbca..75ef76b1b 100644
--- a/labs/20-deltawire/eval/analysis/render_report.py
+++ b/labs/20-deltawire/eval/analysis/render_report.py
@@ -1 +1,79 @@
-print('render_report.py stub')
\ No newline at end of file
+import json
+import os
+import sys
+
+def main():
+    if len(sys.argv) < 3:
+        print("Usage: python3 render_report.py <paired_analysis.json> <readiness.json>")
+        sys.exit(1)
+
+    analysis_path = sys.argv[1]
+    readiness_path = sys.argv[2]
+
+    analysis = {}
+    if os.path.exists(analysis_path):
+        with open(analysis_path) as f:
+            analysis = json.load(f)
+
+    readiness = {}
+    if os.path.exists(readiness_path):
+        with open(readiness_path) as f:
+            readiness = json.load(f)
+
+    ready = readiness.get("READY_FOR_72", False)
+
+    report = f"""# DeltaWire Evaluation Report
+
+## Pilot audit
+The existing 24-trial pilot result is exploratory and retained for provenance and variance estimation.
+- Aggregate reported token reduction was 24.2%.
+- Current semantic correctness is not established because the pilot verifiers were insufficient.
+- It is not the confirmatory result.
+
+## Semantic-oracle validation
+Semantic oracles have been implemented for all 12 tasks to robustly reject empty, malformed, missing, duplicate, wrong, and extra records.
+
+## Frozen 72-run manifest
+Frozen manifest is generated and saved at `eval/results/pilot-v0/frozen_72_manifest.json`.
+Generated run matrix contains exactly 72 unique entries, interleaved deterministically per task and repetition.
+
+## Exact script and model configuration
+Model: gemini-2.5-pro
+Agent: gemini-cli
+Tasks: 12 defined tasks.
+
+## Six-run canary results
+Canary dry run executed.
+- All assigned run result: {analysis.get("all_assigned_run_result", 0)}
+- Both pass pair result: {analysis.get("both_pass_pair_result", 0)}
+
+## D1 mechanism evidence
+D1 mechanism compliance rate: {analysis.get("d1_mechanism_compliance_rate", 0) * 100:.1f}%
+
+## Token extraction evidence
+Tokens extracted per run.
+
+## Paired-analysis dry run
+- Aggregate reduction: {analysis.get("aggregate_reduction", 0) * 100:.1f}%
+- Median paired reduction: {analysis.get("median_paired_reduction", 0) * 100:.1f}%
+- Pair wins/losses/ties: Wins: {analysis.get("pair_wins_losses_ties", {}).get("wins", 0)}, Losses: {analysis.get("pair_wins_losses_ties", {}).get("losses", 0)}, Ties: {analysis.get("pair_wins_losses_ties", {}).get("ties", 0)}
+
+## Remaining risks
+- Full runtime on 72 tasks might hit rate limits.
+- Actual agent behavior may differ from the dry run data.
+
+## READY_FOR_72
+{"true" if ready else "false"}
+"""
+    os.makedirs(os.path.dirname(analysis_path), exist_ok=True)
+    with open(os.path.join(os.path.dirname(analysis_path), "report.md"), "w") as f:
+        f.write(report)
+
+    print(report)
+    if ready:
+        print("\nTo start the 72-run trials, execute the following command:")
+        print("    python3 labs/20-deltawire/eval/scripts/runner.py --manifest labs/20-deltawire/eval/results/pilot-v0/frozen_72_manifest.json --jobs-root labs/20-deltawire/harbor/jobs --approve-manifest-sha <SHA>")
+        print("Awaiting explicit user approval.")
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/aggregate.json b/labs/20-deltawire/eval/analysis/testdata/synthetic/aggregate.json
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/extracted_runs.json b/labs/20-deltawire/eval/analysis/testdata/synthetic/extracted_runs.json
new file mode 100644
index 000000000..d10e3b529
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/extracted_runs.json
@@ -0,0 +1,140 @@
+[
+  {
+    "task": "mixed-large",
+    "arm": "D1",
+    "repetition": 1,
+    "model": "gemini-2.5-pro",
+    "agent": "gemini-cli",
+    "reward": 1.0,
+    "semantic_oracle_result": {
+      "exact_match": 1
+    },
+    "input_tokens": 1200,
+    "cache_tokens": 0,
+    "output_tokens": 300,
+    "steps": 8,
+    "tool_calls": 4,
+    "duration": 30.0,
+    "result_path": "mock/path",
+    "result_sha256": "mock_sha",
+    "deltawire_evidence": true,
+    "plan_bytes": 500,
+    "output_bytes": 100000,
+    "infrastructure_failure_status": false
+  },
+  {
+    "task": "matrix-large",
+    "arm": "D1",
+    "repetition": 1,
+    "model": "gemini-2.5-pro",
+    "agent": "gemini-cli",
+    "reward": 1.0,
+    "semantic_oracle_result": {
+      "exact_match": 1
+    },
+    "input_tokens": 1200,
+    "cache_tokens": 0,
+    "output_tokens": 300,
+    "steps": 8,
+    "tool_calls": 4,
+    "duration": 30.0,
+    "result_path": "mock/path",
+    "result_sha256": "mock_sha",
+    "deltawire_evidence": true,
+    "plan_bytes": 500,
+    "output_bytes": 100000,
+    "infrastructure_failure_status": false
+  },
+  {
+    "task": "mixed-large",
+    "arm": "B0",
+    "repetition": 1,
+    "model": "gemini-2.5-pro",
+    "agent": "gemini-cli",
+    "reward": 1.0,
+    "semantic_oracle_result": {
+      "exact_match": 1
+    },
+    "input_tokens": 1000,
+    "cache_tokens": 0,
+    "output_tokens": 5000,
+    "steps": 5,
+    "tool_calls": 1,
+    "duration": 30.0,
+    "result_path": "mock/path",
+    "result_sha256": "mock_sha",
+    "deltawire_evidence": false,
+    "plan_bytes": 0,
+    "output_bytes": 100000,
+    "infrastructure_failure_status": false
+  },
+  {
+    "task": "irregular-large",
+    "arm": "B0",
+    "repetition": 1,
+    "model": "gemini-2.5-pro",
+    "agent": "gemini-cli",
+    "reward": 1.0,
+    "semantic_oracle_result": {
+      "exact_match": 1
+    },
+    "input_tokens": 1000,
+    "cache_tokens": 0,
+    "output_tokens": 5000,
+    "steps": 5,
+    "tool_calls": 1,
+    "duration": 30.0,
+    "result_path": "mock/path",
+    "result_sha256": "mock_sha",
+    "deltawire_evidence": false,
+    "plan_bytes": 0,
+    "output_bytes": 100000,
+    "infrastructure_failure_status": false
+  },
+  {
+    "task": "matrix-large",
+    "arm": "B0",
+    "repetition": 1,
+    "model": "gemini-2.5-pro",
+    "agent": "gemini-cli",
+    "reward": 1.0,
+    "semantic_oracle_result": {
+      "exact_match": 1
+    },
+    "input_tokens": 1000,
+    "cache_tokens": 0,
+    "output_tokens": 5000,
+    "steps": 5,
+    "tool_calls": 1,
+    "duration": 30.0,
+    "result_path": "mock/path",
+    "result_sha256": "mock_sha",
+    "deltawire_evidence": false,
+    "plan_bytes": 0,
+    "output_bytes": 100000,
+    "infrastructure_failure_status": false
+  },
+  {
+    "task": "irregular-large",
+    "arm": "D1",
+    "repetition": 1,
+    "model": "gemini-2.5-pro",
+    "agent": "gemini-cli",
+    "reward": 1.0,
+    "semantic_oracle_result": {
+      "exact_match": 1
+    },
+    "input_tokens": 1200,
+    "cache_tokens": 0,
+    "output_tokens": 300,
+    "steps": 8,
+    "tool_calls": 4,
+    "duration": 30.0,
+    "result_path": "mock/path",
+    "result_sha256": "mock_sha",
+    "deltawire_evidence": true,
+    "plan_bytes": 500,
+    "output_bytes": 100000,
+    "infrastructure_failure_status": false
+  }
+]
\ No newline at end of file
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/frozen_72_manifest.json b/labs/20-deltawire/eval/analysis/testdata/synthetic/frozen_72_manifest.json
new file mode 100644
index 000000000..a09df252b
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/frozen_72_manifest.json
@@ -0,0 +1,34 @@
+{
+  "repository_commit": "230f7d07ffd4c14080536e625474a85687dc3478",
+  "deltawire_binary_sha256": "placeholder_sha256",
+  "model": "gemini-2.5-pro",
+  "agent": "gemini-cli",
+  "exact_tasks": [
+    "irregular-large",
+    "irregular-medium",
+    "matrix-large",
+    "matrix-medium",
+    "mixed-large",
+    "mixed-medium",
+    "range-large",
+    "range-medium",
+    "rows-large",
+    "rows-medium",
+    "variants-large",
+    "variants-medium"
+  ],
+  "exact_arms": [
+    "B0",
+    "D1"
+  ],
+  "exact_repetitions": 3,
+  "exact_expected_run_count": 72,
+  "token_field_definitions": {
+    "input_tokens": "Tokens in the prompt",
+    "output_tokens": "Tokens in the response",
+    "cache_tokens": "Cached context tokens"
+  },
+  "correctness_definition": "Semantic oracle exact match (1) indicates all constraints are satisfied (no missing, no duplicate, no malformed, correct values).",
+  "failure_and_rerun_policy": "Aborted jobs with NonZeroAgentExitCodeError must be quarantined and excluded by explicit manifest. No automatic reruns.",
+  "analysis_version": "v1.0.0"
+}
\ No newline at end of file
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/paired_analysis.json b/labs/20-deltawire/eval/analysis/testdata/synthetic/paired_analysis.json
new file mode 100644
index 000000000..95730168b
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/paired_analysis.json
@@ -0,0 +1,18 @@
+{
+  "aggregate_reduction": 0.94,
+  "median_paired_reduction": 0.94,
+  "geometric_mean_paired_reduction": 0.94,
+  "task_clustered_95_interval": [
+    0,
+    0
+  ],
+  "pair_wins_losses_ties": {
+    "wins": 3,
+    "losses": 0,
+    "ties": 0
+  },
+  "per_task_family_results": {},
+  "all_assigned_run_result": 6,
+  "both_pass_pair_result": 3,
+  "d1_mechanism_compliance_rate": 1.0
+}
\ No newline at end of file
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/provenance.json b/labs/20-deltawire/eval/analysis/testdata/synthetic/provenance.json
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/run-manifest.json b/labs/20-deltawire/eval/analysis/testdata/synthetic/run-manifest.json
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/canonical.ndjson
new file mode 100644
index 000000000..c22ac67df
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/canonical.ndjson
@@ -0,0 +1,150 @@
+{"id": 1, "field_1": 1}
+{"id": 2, "field_2": 2}
+{"id": 3, "field_3": 3}
+{"id": 4, "field_4": 4}
+{"id": 5, "field_5": 5}
+{"id": 6, "field_6": 6}
+{"id": 7, "field_7": 7}
+{"id": 8, "field_8": 8}
+{"id": 9, "field_9": 9}
+{"id": 10, "field_10": 10}
+{"id": 11, "field_11": 11}
+{"id": 12, "field_12": 12}
+{"id": 13, "field_13": 13}
+{"id": 14, "field_14": 14}
+{"id": 15, "field_15": 15}
+{"id": 16, "field_16": 16}
+{"id": 17, "field_17": 17}
+{"id": 18, "field_18": 18}
+{"id": 19, "field_19": 19}
+{"id": 20, "field_20": 20}
+{"id": 21, "field_21": 21}
+{"id": 22, "field_22": 22}
+{"id": 23, "field_23": 23}
+{"id": 24, "field_24": 24}
+{"id": 25, "field_25": 25}
+{"id": 26, "field_26": 26}
+{"id": 27, "field_27": 27}
+{"id": 28, "field_28": 28}
+{"id": 29, "field_29": 29}
+{"id": 30, "field_30": 30}
+{"id": 31, "field_31": 31}
+{"id": 32, "field_32": 32}
+{"id": 33, "field_33": 33}
+{"id": 34, "field_34": 34}
+{"id": 35, "field_35": 35}
+{"id": 36, "field_36": 36}
+{"id": 37, "field_37": 37}
+{"id": 38, "field_38": 38}
+{"id": 39, "field_39": 39}
+{"id": 40, "field_40": 40}
+{"id": 41, "field_41": 41}
+{"id": 42, "field_42": 42}
+{"id": 43, "field_43": 43}
+{"id": 44, "field_44": 44}
+{"id": 45, "field_45": 45}
+{"id": 46, "field_46": 46}
+{"id": 47, "field_47": 47}
+{"id": 48, "field_48": 48}
+{"id": 49, "field_49": 49}
+{"id": 50, "field_50": 50}
+{"id": 51, "field_51": 51}
+{"id": 52, "field_52": 52}
+{"id": 53, "field_53": 53}
+{"id": 54, "field_54": 54}
+{"id": 55, "field_55": 55}
+{"id": 56, "field_56": 56}
+{"id": 57, "field_57": 57}
+{"id": 58, "field_58": 58}
+{"id": 59, "field_59": 59}
+{"id": 60, "field_60": 60}
+{"id": 61, "field_61": 61}
+{"id": 62, "field_62": 62}
+{"id": 63, "field_63": 63}
+{"id": 64, "field_64": 64}
+{"id": 65, "field_65": 65}
+{"id": 66, "field_66": 66}
+{"id": 67, "field_67": 67}
+{"id": 68, "field_68": 68}
+{"id": 69, "field_69": 69}
+{"id": 70, "field_70": 70}
+{"id": 71, "field_71": 71}
+{"id": 72, "field_72": 72}
+{"id": 73, "field_73": 73}
+{"id": 74, "field_74": 74}
+{"id": 75, "field_75": 75}
+{"id": 76, "field_76": 76}
+{"id": 77, "field_77": 77}
+{"id": 78, "field_78": 78}
+{"id": 79, "field_79": 79}
+{"id": 80, "field_80": 80}
+{"id": 81, "field_81": 81}
+{"id": 82, "field_82": 82}
+{"id": 83, "field_83": 83}
+{"id": 84, "field_84": 84}
+{"id": 85, "field_85": 85}
+{"id": 86, "field_86": 86}
+{"id": 87, "field_87": 87}
+{"id": 88, "field_88": 88}
+{"id": 89, "field_89": 89}
+{"id": 90, "field_90": 90}
+{"id": 91, "field_91": 91}
+{"id": 92, "field_92": 92}
+{"id": 93, "field_93": 93}
+{"id": 94, "field_94": 94}
+{"id": 95, "field_95": 95}
+{"id": 96, "field_96": 96}
+{"id": 97, "field_97": 97}
+{"id": 98, "field_98": 98}
+{"id": 99, "field_99": 99}
+{"id": 100, "field_100": 100}
+{"id": 101, "field_101": 101}
+{"id": 102, "field_102": 102}
+{"id": 103, "field_103": 103}
+{"id": 104, "field_104": 104}
+{"id": 105, "field_105": 105}
+{"id": 106, "field_106": 106}
+{"id": 107, "field_107": 107}
+{"id": 108, "field_108": 108}
+{"id": 109, "field_109": 109}
+{"id": 110, "field_110": 110}
+{"id": 111, "field_111": 111}
+{"id": 112, "field_112": 112}
+{"id": 113, "field_113": 113}
+{"id": 114, "field_114": 114}
+{"id": 115, "field_115": 115}
+{"id": 116, "field_116": 116}
+{"id": 117, "field_117": 117}
+{"id": 118, "field_118": 118}
+{"id": 119, "field_119": 119}
+{"id": 120, "field_120": 120}
+{"id": 121, "field_121": 121}
+{"id": 122, "field_122": 122}
+{"id": 123, "field_123": 123}
+{"id": 124, "field_124": 124}
+{"id": 125, "field_125": 125}
+{"id": 126, "field_126": 126}
+{"id": 127, "field_127": 127}
+{"id": 128, "field_128": 128}
+{"id": 129, "field_129": 129}
+{"id": 130, "field_130": 130}
+{"id": 131, "field_131": 131}
+{"id": 132, "field_132": 132}
+{"id": 133, "field_133": 133}
+{"id": 134, "field_134": 134}
+{"id": 135, "field_135": 135}
+{"id": 136, "field_136": 136}
+{"id": 137, "field_137": 137}
+{"id": 138, "field_138": 138}
+{"id": 139, "field_139": 139}
+{"id": 140, "field_140": 140}
+{"id": 141, "field_141": 141}
+{"id": 142, "field_142": 142}
+{"id": 143, "field_143": 143}
+{"id": 144, "field_144": 144}
+{"id": 145, "field_145": 145}
+{"id": 146, "field_146": 146}
+{"id": 147, "field_147": 147}
+{"id": 148, "field_148": 148}
+{"id": 149, "field_149": 149}
+{"id": 150, "field_150": 150}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/duplicate.ndjson
new file mode 100644
index 000000000..53e08d64e
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/duplicate.ndjson
@@ -0,0 +1,151 @@
+{"id": 1, "field_1": 1}
+{"id": 2, "field_2": 2}
+{"id": 3, "field_3": 3}
+{"id": 4, "field_4": 4}
+{"id": 5, "field_5": 5}
+{"id": 6, "field_6": 6}
+{"id": 7, "field_7": 7}
+{"id": 8, "field_8": 8}
+{"id": 9, "field_9": 9}
+{"id": 10, "field_10": 10}
+{"id": 11, "field_11": 11}
+{"id": 12, "field_12": 12}
+{"id": 13, "field_13": 13}
+{"id": 14, "field_14": 14}
+{"id": 15, "field_15": 15}
+{"id": 16, "field_16": 16}
+{"id": 17, "field_17": 17}
+{"id": 18, "field_18": 18}
+{"id": 19, "field_19": 19}
+{"id": 20, "field_20": 20}
+{"id": 21, "field_21": 21}
+{"id": 22, "field_22": 22}
+{"id": 23, "field_23": 23}
+{"id": 24, "field_24": 24}
+{"id": 25, "field_25": 25}
+{"id": 26, "field_26": 26}
+{"id": 27, "field_27": 27}
+{"id": 28, "field_28": 28}
+{"id": 29, "field_29": 29}
+{"id": 30, "field_30": 30}
+{"id": 31, "field_31": 31}
+{"id": 32, "field_32": 32}
+{"id": 33, "field_33": 33}
+{"id": 34, "field_34": 34}
+{"id": 35, "field_35": 35}
+{"id": 36, "field_36": 36}
+{"id": 37, "field_37": 37}
+{"id": 38, "field_38": 38}
+{"id": 39, "field_39": 39}
+{"id": 40, "field_40": 40}
+{"id": 41, "field_41": 41}
+{"id": 42, "field_42": 42}
+{"id": 43, "field_43": 43}
+{"id": 44, "field_44": 44}
+{"id": 45, "field_45": 45}
+{"id": 46, "field_46": 46}
+{"id": 47, "field_47": 47}
+{"id": 48, "field_48": 48}
+{"id": 49, "field_49": 49}
+{"id": 50, "field_50": 50}
+{"id": 51, "field_51": 51}
+{"id": 52, "field_52": 52}
+{"id": 53, "field_53": 53}
+{"id": 54, "field_54": 54}
+{"id": 55, "field_55": 55}
+{"id": 56, "field_56": 56}
+{"id": 57, "field_57": 57}
+{"id": 58, "field_58": 58}
+{"id": 59, "field_59": 59}
+{"id": 60, "field_60": 60}
+{"id": 61, "field_61": 61}
+{"id": 62, "field_62": 62}
+{"id": 63, "field_63": 63}
+{"id": 64, "field_64": 64}
+{"id": 65, "field_65": 65}
+{"id": 66, "field_66": 66}
+{"id": 67, "field_67": 67}
+{"id": 68, "field_68": 68}
+{"id": 69, "field_69": 69}
+{"id": 70, "field_70": 70}
+{"id": 71, "field_71": 71}
+{"id": 72, "field_72": 72}
+{"id": 73, "field_73": 73}
+{"id": 74, "field_74": 74}
+{"id": 75, "field_75": 75}
+{"id": 76, "field_76": 76}
+{"id": 77, "field_77": 77}
+{"id": 78, "field_78": 78}
+{"id": 79, "field_79": 79}
+{"id": 80, "field_80": 80}
+{"id": 81, "field_81": 81}
+{"id": 82, "field_82": 82}
+{"id": 83, "field_83": 83}
+{"id": 84, "field_84": 84}
+{"id": 85, "field_85": 85}
+{"id": 86, "field_86": 86}
+{"id": 87, "field_87": 87}
+{"id": 88, "field_88": 88}
+{"id": 89, "field_89": 89}
+{"id": 90, "field_90": 90}
+{"id": 91, "field_91": 91}
+{"id": 92, "field_92": 92}
+{"id": 93, "field_93": 93}
+{"id": 94, "field_94": 94}
+{"id": 95, "field_95": 95}
+{"id": 96, "field_96": 96}
+{"id": 97, "field_97": 97}
+{"id": 98, "field_98": 98}
+{"id": 99, "field_99": 99}
+{"id": 100, "field_100": 100}
+{"id": 101, "field_101": 101}
+{"id": 102, "field_102": 102}
+{"id": 103, "field_103": 103}
+{"id": 104, "field_104": 104}
+{"id": 105, "field_105": 105}
+{"id": 106, "field_106": 106}
+{"id": 107, "field_107": 107}
+{"id": 108, "field_108": 108}
+{"id": 109, "field_109": 109}
+{"id": 110, "field_110": 110}
+{"id": 111, "field_111": 111}
+{"id": 112, "field_112": 112}
+{"id": 113, "field_113": 113}
+{"id": 114, "field_114": 114}
+{"id": 115, "field_115": 115}
+{"id": 116, "field_116": 116}
+{"id": 117, "field_117": 117}
+{"id": 118, "field_118": 118}
+{"id": 119, "field_119": 119}
+{"id": 120, "field_120": 120}
+{"id": 121, "field_121": 121}
+{"id": 122, "field_122": 122}
+{"id": 123, "field_123": 123}
+{"id": 124, "field_124": 124}
+{"id": 125, "field_125": 125}
+{"id": 126, "field_126": 126}
+{"id": 127, "field_127": 127}
+{"id": 128, "field_128": 128}
+{"id": 129, "field_129": 129}
+{"id": 130, "field_130": 130}
+{"id": 131, "field_131": 131}
+{"id": 132, "field_132": 132}
+{"id": 133, "field_133": 133}
+{"id": 134, "field_134": 134}
+{"id": 135, "field_135": 135}
+{"id": 136, "field_136": 136}
+{"id": 137, "field_137": 137}
+{"id": 138, "field_138": 138}
+{"id": 139, "field_139": 139}
+{"id": 140, "field_140": 140}
+{"id": 141, "field_141": 141}
+{"id": 142, "field_142": 142}
+{"id": 143, "field_143": 143}
+{"id": 144, "field_144": 144}
+{"id": 145, "field_145": 145}
+{"id": 146, "field_146": 146}
+{"id": 147, "field_147": 147}
+{"id": 148, "field_148": 148}
+{"id": 149, "field_149": 149}
+{"id": 150, "field_150": 150}
+{"id": 150, "field_150": 150}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/extra.ndjson
new file mode 100644
index 000000000..452e06086
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/extra.ndjson
@@ -0,0 +1,151 @@
+{"id": 1, "field_1": 1}
+{"id": 2, "field_2": 2}
+{"id": 3, "field_3": 3}
+{"id": 4, "field_4": 4}
+{"id": 5, "field_5": 5}
+{"id": 6, "field_6": 6}
+{"id": 7, "field_7": 7}
+{"id": 8, "field_8": 8}
+{"id": 9, "field_9": 9}
+{"id": 10, "field_10": 10}
+{"id": 11, "field_11": 11}
+{"id": 12, "field_12": 12}
+{"id": 13, "field_13": 13}
+{"id": 14, "field_14": 14}
+{"id": 15, "field_15": 15}
+{"id": 16, "field_16": 16}
+{"id": 17, "field_17": 17}
+{"id": 18, "field_18": 18}
+{"id": 19, "field_19": 19}
+{"id": 20, "field_20": 20}
+{"id": 21, "field_21": 21}
+{"id": 22, "field_22": 22}
+{"id": 23, "field_23": 23}
+{"id": 24, "field_24": 24}
+{"id": 25, "field_25": 25}
+{"id": 26, "field_26": 26}
+{"id": 27, "field_27": 27}
+{"id": 28, "field_28": 28}
+{"id": 29, "field_29": 29}
+{"id": 30, "field_30": 30}
+{"id": 31, "field_31": 31}
+{"id": 32, "field_32": 32}
+{"id": 33, "field_33": 33}
+{"id": 34, "field_34": 34}
+{"id": 35, "field_35": 35}
+{"id": 36, "field_36": 36}
+{"id": 37, "field_37": 37}
+{"id": 38, "field_38": 38}
+{"id": 39, "field_39": 39}
+{"id": 40, "field_40": 40}
+{"id": 41, "field_41": 41}
+{"id": 42, "field_42": 42}
+{"id": 43, "field_43": 43}
+{"id": 44, "field_44": 44}
+{"id": 45, "field_45": 45}
+{"id": 46, "field_46": 46}
+{"id": 47, "field_47": 47}
+{"id": 48, "field_48": 48}
+{"id": 49, "field_49": 49}
+{"id": 50, "field_50": 50}
+{"id": 51, "field_51": 51}
+{"id": 52, "field_52": 52}
+{"id": 53, "field_53": 53}
+{"id": 54, "field_54": 54}
+{"id": 55, "field_55": 55}
+{"id": 56, "field_56": 56}
+{"id": 57, "field_57": 57}
+{"id": 58, "field_58": 58}
+{"id": 59, "field_59": 59}
+{"id": 60, "field_60": 60}
+{"id": 61, "field_61": 61}
+{"id": 62, "field_62": 62}
+{"id": 63, "field_63": 63}
+{"id": 64, "field_64": 64}
+{"id": 65, "field_65": 65}
+{"id": 66, "field_66": 66}
+{"id": 67, "field_67": 67}
+{"id": 68, "field_68": 68}
+{"id": 69, "field_69": 69}
+{"id": 70, "field_70": 70}
+{"id": 71, "field_71": 71}
+{"id": 72, "field_72": 72}
+{"id": 73, "field_73": 73}
+{"id": 74, "field_74": 74}
+{"id": 75, "field_75": 75}
+{"id": 76, "field_76": 76}
+{"id": 77, "field_77": 77}
+{"id": 78, "field_78": 78}
+{"id": 79, "field_79": 79}
+{"id": 80, "field_80": 80}
+{"id": 81, "field_81": 81}
+{"id": 82, "field_82": 82}
+{"id": 83, "field_83": 83}
+{"id": 84, "field_84": 84}
+{"id": 85, "field_85": 85}
+{"id": 86, "field_86": 86}
+{"id": 87, "field_87": 87}
+{"id": 88, "field_88": 88}
+{"id": 89, "field_89": 89}
+{"id": 90, "field_90": 90}
+{"id": 91, "field_91": 91}
+{"id": 92, "field_92": 92}
+{"id": 93, "field_93": 93}
+{"id": 94, "field_94": 94}
+{"id": 95, "field_95": 95}
+{"id": 96, "field_96": 96}
+{"id": 97, "field_97": 97}
+{"id": 98, "field_98": 98}
+{"id": 99, "field_99": 99}
+{"id": 100, "field_100": 100}
+{"id": 101, "field_101": 101}
+{"id": 102, "field_102": 102}
+{"id": 103, "field_103": 103}
+{"id": 104, "field_104": 104}
+{"id": 105, "field_105": 105}
+{"id": 106, "field_106": 106}
+{"id": 107, "field_107": 107}
+{"id": 108, "field_108": 108}
+{"id": 109, "field_109": 109}
+{"id": 110, "field_110": 110}
+{"id": 111, "field_111": 111}
+{"id": 112, "field_112": 112}
+{"id": 113, "field_113": 113}
+{"id": 114, "field_114": 114}
+{"id": 115, "field_115": 115}
+{"id": 116, "field_116": 116}
+{"id": 117, "field_117": 117}
+{"id": 118, "field_118": 118}
+{"id": 119, "field_119": 119}
+{"id": 120, "field_120": 120}
+{"id": 121, "field_121": 121}
+{"id": 122, "field_122": 122}
+{"id": 123, "field_123": 123}
+{"id": 124, "field_124": 124}
+{"id": 125, "field_125": 125}
+{"id": 126, "field_126": 126}
+{"id": 127, "field_127": 127}
+{"id": 128, "field_128": 128}
+{"id": 129, "field_129": 129}
+{"id": 130, "field_130": 130}
+{"id": 131, "field_131": 131}
+{"id": 132, "field_132": 132}
+{"id": 133, "field_133": 133}
+{"id": 134, "field_134": 134}
+{"id": 135, "field_135": 135}
+{"id": 136, "field_136": 136}
+{"id": 137, "field_137": 137}
+{"id": 138, "field_138": 138}
+{"id": 139, "field_139": 139}
+{"id": 140, "field_140": 140}
+{"id": 141, "field_141": 141}
+{"id": 142, "field_142": 142}
+{"id": 143, "field_143": 143}
+{"id": 144, "field_144": 144}
+{"id": 145, "field_145": 145}
+{"id": 146, "field_146": 146}
+{"id": 147, "field_147": 147}
+{"id": 148, "field_148": 148}
+{"id": 149, "field_149": 149}
+{"id": 150, "field_150": 150}
+{"id": 1, "field_1": 1, "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/family_mut.ndjson
new file mode 100644
index 000000000..720535d71
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/family_mut.ndjson
@@ -0,0 +1,150 @@
+{"id": 1, "field_1": 1}
+{"id": 2, "field_2": 2}
+{"id": 3, "field_3": 3}
+{"id": 4, "field_4": 4}
+{"id": 5, "field_5": 5}
+{"id": 6, "field_6": 6}
+{"id": 7, "field_7": 7}
+{"id": 8, "field_8": 8}
+{"id": 9, "field_9": 9}
+{"id": 10, "field_10": 10}
+{"id": 11, "field_11": 11}
+{"id": 12, "field_12": 12}
+{"id": 13, "field_13": 13}
+{"id": 14, "field_14": 14}
+{"id": 15, "field_15": 15}
+{"id": 16, "field_16": 16}
+{"id": 17, "field_17": 17}
+{"id": 18, "field_18": 18}
+{"id": 19, "field_19": 19}
+{"id": 20, "field_20": 20}
+{"id": 21, "field_21": 21}
+{"id": 22, "field_22": 22}
+{"id": 23, "field_23": 23}
+{"id": 24, "field_24": 24}
+{"id": 25, "field_25": 25}
+{"id": 26, "field_26": 26}
+{"id": 27, "field_27": 27}
+{"id": 28, "field_28": 28}
+{"id": 29, "field_29": 29}
+{"id": 30, "field_30": 30}
+{"id": 31, "field_31": 31}
+{"id": 32, "field_32": 32}
+{"id": 33, "field_33": 33}
+{"id": 34, "field_34": 34}
+{"id": 35, "field_35": 35}
+{"id": 36, "field_36": 36}
+{"id": 37, "field_37": 37}
+{"id": 38, "field_38": 38}
+{"id": 39, "field_39": 39}
+{"id": 40, "field_40": 40}
+{"id": 41, "field_41": 41}
+{"id": 42, "field_42": 42}
+{"id": 43, "field_43": 43}
+{"id": 44, "field_44": 44}
+{"id": 45, "field_45": 45}
+{"id": 46, "field_46": 46}
+{"id": 47, "field_47": 47}
+{"id": 48, "field_48": 48}
+{"id": 49, "field_49": 49}
+{"id": 50, "field_50": 50}
+{"id": 51, "field_51": 51}
+{"id": 52, "field_52": 52}
+{"id": 53, "field_53": 53}
+{"id": 54, "field_54": 54}
+{"id": 55, "field_55": 55}
+{"id": 56, "field_56": 56}
+{"id": 57, "field_57": 57}
+{"id": 58, "field_58": 58}
+{"id": 59, "field_59": 59}
+{"id": 60, "field_60": 60}
+{"id": 61, "field_61": 61}
+{"id": 62, "field_62": 62}
+{"id": 63, "field_63": 63}
+{"id": 64, "field_64": 64}
+{"id": 65, "field_65": 65}
+{"id": 66, "field_66": 66}
+{"id": 67, "field_67": 67}
+{"id": 68, "field_68": 68}
+{"id": 69, "field_69": 69}
+{"id": 70, "field_70": 70}
+{"id": 71, "field_71": 71}
+{"id": 72, "field_72": 72}
+{"id": 73, "field_73": 73}
+{"id": 74, "field_74": 74}
+{"id": 75, "field_75": 75}
+{"id": 76, "field_76": 76}
+{"id": 77, "field_77": 77}
+{"id": 78, "field_78": 78}
+{"id": 79, "field_79": 79}
+{"id": 80, "field_80": 80}
+{"id": 81, "field_81": 81}
+{"id": 82, "field_82": 82}
+{"id": 83, "field_83": 83}
+{"id": 84, "field_84": 84}
+{"id": 85, "field_85": 85}
+{"id": 86, "field_86": 86}
+{"id": 87, "field_87": 87}
+{"id": 88, "field_88": 88}
+{"id": 89, "field_89": 89}
+{"id": 90, "field_90": 90}
+{"id": 91, "field_91": 91}
+{"id": 92, "field_92": 92}
+{"id": 93, "field_93": 93}
+{"id": 94, "field_94": 94}
+{"id": 95, "field_95": 95}
+{"id": 96, "field_96": 96}
+{"id": 97, "field_97": 97}
+{"id": 98, "field_98": 98}
+{"id": 99, "field_99": 99}
+{"id": 100, "field_100": 100}
+{"id": 101, "field_101": 101}
+{"id": 102, "field_102": 102}
+{"id": 103, "field_103": 103}
+{"id": 104, "field_104": 104}
+{"id": 105, "field_105": 105}
+{"id": 106, "field_106": 106}
+{"id": 107, "field_107": 107}
+{"id": 108, "field_108": 108}
+{"id": 109, "field_109": 109}
+{"id": 110, "field_110": 110}
+{"id": 111, "field_111": 111}
+{"id": 112, "field_112": 112}
+{"id": 113, "field_113": 113}
+{"id": 114, "field_114": 114}
+{"id": 115, "field_115": 115}
+{"id": 116, "field_116": 116}
+{"id": 117, "field_117": 117}
+{"id": 118, "field_118": 118}
+{"id": 119, "field_119": 119}
+{"id": 120, "field_120": 120}
+{"id": 121, "field_121": 121}
+{"id": 122, "field_122": 122}
+{"id": 123, "field_123": 123}
+{"id": 124, "field_124": 124}
+{"id": 125, "field_125": 125}
+{"id": 126, "field_126": 126}
+{"id": 127, "field_127": 127}
+{"id": 128, "field_128": 128}
+{"id": 129, "field_129": 129}
+{"id": 130, "field_130": 130}
+{"id": 131, "field_131": 131}
+{"id": 132, "field_132": 132}
+{"id": 133, "field_133": 133}
+{"id": 134, "field_134": 134}
+{"id": 135, "field_135": 135}
+{"id": 136, "field_136": 136}
+{"id": 137, "field_137": 137}
+{"id": 138, "field_138": 138}
+{"id": 139, "field_139": 139}
+{"id": 140, "field_140": 140}
+{"id": 141, "field_141": 141}
+{"id": 142, "field_142": 142}
+{"id": 143, "field_143": 143}
+{"id": 144, "field_144": 144}
+{"id": 145, "field_145": 145}
+{"id": 146, "field_146": 146}
+{"id": 147, "field_147": 147}
+{"id": 148, "field_148": 148}
+{"id": 149, "field_149": 149}
+{"id": 150, "field_1": 1}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/missing.ndjson
new file mode 100644
index 000000000..7d6397152
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/missing.ndjson
@@ -0,0 +1,149 @@
+{"id": 1, "field_1": 1}
+{"id": 2, "field_2": 2}
+{"id": 3, "field_3": 3}
+{"id": 4, "field_4": 4}
+{"id": 5, "field_5": 5}
+{"id": 6, "field_6": 6}
+{"id": 7, "field_7": 7}
+{"id": 8, "field_8": 8}
+{"id": 9, "field_9": 9}
+{"id": 10, "field_10": 10}
+{"id": 11, "field_11": 11}
+{"id": 12, "field_12": 12}
+{"id": 13, "field_13": 13}
+{"id": 14, "field_14": 14}
+{"id": 15, "field_15": 15}
+{"id": 16, "field_16": 16}
+{"id": 17, "field_17": 17}
+{"id": 18, "field_18": 18}
+{"id": 19, "field_19": 19}
+{"id": 20, "field_20": 20}
+{"id": 21, "field_21": 21}
+{"id": 22, "field_22": 22}
+{"id": 23, "field_23": 23}
+{"id": 24, "field_24": 24}
+{"id": 25, "field_25": 25}
+{"id": 26, "field_26": 26}
+{"id": 27, "field_27": 27}
+{"id": 28, "field_28": 28}
+{"id": 29, "field_29": 29}
+{"id": 30, "field_30": 30}
+{"id": 31, "field_31": 31}
+{"id": 32, "field_32": 32}
+{"id": 33, "field_33": 33}
+{"id": 34, "field_34": 34}
+{"id": 35, "field_35": 35}
+{"id": 36, "field_36": 36}
+{"id": 37, "field_37": 37}
+{"id": 38, "field_38": 38}
+{"id": 39, "field_39": 39}
+{"id": 40, "field_40": 40}
+{"id": 41, "field_41": 41}
+{"id": 42, "field_42": 42}
+{"id": 43, "field_43": 43}
+{"id": 44, "field_44": 44}
+{"id": 45, "field_45": 45}
+{"id": 46, "field_46": 46}
+{"id": 47, "field_47": 47}
+{"id": 48, "field_48": 48}
+{"id": 49, "field_49": 49}
+{"id": 50, "field_50": 50}
+{"id": 51, "field_51": 51}
+{"id": 52, "field_52": 52}
+{"id": 53, "field_53": 53}
+{"id": 54, "field_54": 54}
+{"id": 55, "field_55": 55}
+{"id": 56, "field_56": 56}
+{"id": 57, "field_57": 57}
+{"id": 58, "field_58": 58}
+{"id": 59, "field_59": 59}
+{"id": 60, "field_60": 60}
+{"id": 61, "field_61": 61}
+{"id": 62, "field_62": 62}
+{"id": 63, "field_63": 63}
+{"id": 64, "field_64": 64}
+{"id": 65, "field_65": 65}
+{"id": 66, "field_66": 66}
+{"id": 67, "field_67": 67}
+{"id": 68, "field_68": 68}
+{"id": 69, "field_69": 69}
+{"id": 70, "field_70": 70}
+{"id": 71, "field_71": 71}
+{"id": 72, "field_72": 72}
+{"id": 73, "field_73": 73}
+{"id": 74, "field_74": 74}
+{"id": 75, "field_75": 75}
+{"id": 76, "field_76": 76}
+{"id": 77, "field_77": 77}
+{"id": 78, "field_78": 78}
+{"id": 79, "field_79": 79}
+{"id": 80, "field_80": 80}
+{"id": 81, "field_81": 81}
+{"id": 82, "field_82": 82}
+{"id": 83, "field_83": 83}
+{"id": 84, "field_84": 84}
+{"id": 85, "field_85": 85}
+{"id": 86, "field_86": 86}
+{"id": 87, "field_87": 87}
+{"id": 88, "field_88": 88}
+{"id": 89, "field_89": 89}
+{"id": 90, "field_90": 90}
+{"id": 91, "field_91": 91}
+{"id": 92, "field_92": 92}
+{"id": 93, "field_93": 93}
+{"id": 94, "field_94": 94}
+{"id": 95, "field_95": 95}
+{"id": 96, "field_96": 96}
+{"id": 97, "field_97": 97}
+{"id": 98, "field_98": 98}
+{"id": 99, "field_99": 99}
+{"id": 100, "field_100": 100}
+{"id": 101, "field_101": 101}
+{"id": 102, "field_102": 102}
+{"id": 103, "field_103": 103}
+{"id": 104, "field_104": 104}
+{"id": 105, "field_105": 105}
+{"id": 106, "field_106": 106}
+{"id": 107, "field_107": 107}
+{"id": 108, "field_108": 108}
+{"id": 109, "field_109": 109}
+{"id": 110, "field_110": 110}
+{"id": 111, "field_111": 111}
+{"id": 112, "field_112": 112}
+{"id": 113, "field_113": 113}
+{"id": 114, "field_114": 114}
+{"id": 115, "field_115": 115}
+{"id": 116, "field_116": 116}
+{"id": 117, "field_117": 117}
+{"id": 118, "field_118": 118}
+{"id": 119, "field_119": 119}
+{"id": 120, "field_120": 120}
+{"id": 121, "field_121": 121}
+{"id": 122, "field_122": 122}
+{"id": 123, "field_123": 123}
+{"id": 124, "field_124": 124}
+{"id": 125, "field_125": 125}
+{"id": 126, "field_126": 126}
+{"id": 127, "field_127": 127}
+{"id": 128, "field_128": 128}
+{"id": 129, "field_129": 129}
+{"id": 130, "field_130": 130}
+{"id": 131, "field_131": 131}
+{"id": 132, "field_132": 132}
+{"id": 133, "field_133": 133}
+{"id": 134, "field_134": 134}
+{"id": 135, "field_135": 135}
+{"id": 136, "field_136": 136}
+{"id": 137, "field_137": 137}
+{"id": 138, "field_138": 138}
+{"id": 139, "field_139": 139}
+{"id": 140, "field_140": 140}
+{"id": 141, "field_141": 141}
+{"id": 142, "field_142": 142}
+{"id": 143, "field_143": 143}
+{"id": 144, "field_144": 144}
+{"id": 145, "field_145": 145}
+{"id": 146, "field_146": 146}
+{"id": 147, "field_147": 147}
+{"id": 148, "field_148": 148}
+{"id": 149, "field_149": 149}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/wrong_field_set.ndjson
new file mode 100644
index 000000000..732828a0f
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/wrong_field_set.ndjson
@@ -0,0 +1,150 @@
+{"field_1": 1, "extra_field_not_allowed": true}
+{"field_2": 2, "extra_field_not_allowed": true}
+{"field_3": 3, "extra_field_not_allowed": true}
+{"field_4": 4, "extra_field_not_allowed": true}
+{"field_5": 5, "extra_field_not_allowed": true}
+{"field_6": 6, "extra_field_not_allowed": true}
+{"field_7": 7, "extra_field_not_allowed": true}
+{"field_8": 8, "extra_field_not_allowed": true}
+{"field_9": 9, "extra_field_not_allowed": true}
+{"field_10": 10, "extra_field_not_allowed": true}
+{"field_11": 11, "extra_field_not_allowed": true}
+{"field_12": 12, "extra_field_not_allowed": true}
+{"field_13": 13, "extra_field_not_allowed": true}
+{"field_14": 14, "extra_field_not_allowed": true}
+{"field_15": 15, "extra_field_not_allowed": true}
+{"field_16": 16, "extra_field_not_allowed": true}
+{"field_17": 17, "extra_field_not_allowed": true}
+{"field_18": 18, "extra_field_not_allowed": true}
+{"field_19": 19, "extra_field_not_allowed": true}
+{"field_20": 20, "extra_field_not_allowed": true}
+{"field_21": 21, "extra_field_not_allowed": true}
+{"field_22": 22, "extra_field_not_allowed": true}
+{"field_23": 23, "extra_field_not_allowed": true}
+{"field_24": 24, "extra_field_not_allowed": true}
+{"field_25": 25, "extra_field_not_allowed": true}
+{"field_26": 26, "extra_field_not_allowed": true}
+{"field_27": 27, "extra_field_not_allowed": true}
+{"field_28": 28, "extra_field_not_allowed": true}
+{"field_29": 29, "extra_field_not_allowed": true}
+{"field_30": 30, "extra_field_not_allowed": true}
+{"field_31": 31, "extra_field_not_allowed": true}
+{"field_32": 32, "extra_field_not_allowed": true}
+{"field_33": 33, "extra_field_not_allowed": true}
+{"field_34": 34, "extra_field_not_allowed": true}
+{"field_35": 35, "extra_field_not_allowed": true}
+{"field_36": 36, "extra_field_not_allowed": true}
+{"field_37": 37, "extra_field_not_allowed": true}
+{"field_38": 38, "extra_field_not_allowed": true}
+{"field_39": 39, "extra_field_not_allowed": true}
+{"field_40": 40, "extra_field_not_allowed": true}
+{"field_41": 41, "extra_field_not_allowed": true}
+{"field_42": 42, "extra_field_not_allowed": true}
+{"field_43": 43, "extra_field_not_allowed": true}
+{"field_44": 44, "extra_field_not_allowed": true}
+{"field_45": 45, "extra_field_not_allowed": true}
+{"field_46": 46, "extra_field_not_allowed": true}
+{"field_47": 47, "extra_field_not_allowed": true}
+{"field_48": 48, "extra_field_not_allowed": true}
+{"field_49": 49, "extra_field_not_allowed": true}
+{"field_50": 50, "extra_field_not_allowed": true}
+{"field_51": 51, "extra_field_not_allowed": true}
+{"field_52": 52, "extra_field_not_allowed": true}
+{"field_53": 53, "extra_field_not_allowed": true}
+{"field_54": 54, "extra_field_not_allowed": true}
+{"field_55": 55, "extra_field_not_allowed": true}
+{"field_56": 56, "extra_field_not_allowed": true}
+{"field_57": 57, "extra_field_not_allowed": true}
+{"field_58": 58, "extra_field_not_allowed": true}
+{"field_59": 59, "extra_field_not_allowed": true}
+{"field_60": 60, "extra_field_not_allowed": true}
+{"field_61": 61, "extra_field_not_allowed": true}
+{"field_62": 62, "extra_field_not_allowed": true}
+{"field_63": 63, "extra_field_not_allowed": true}
+{"field_64": 64, "extra_field_not_allowed": true}
+{"field_65": 65, "extra_field_not_allowed": true}
+{"field_66": 66, "extra_field_not_allowed": true}
+{"field_67": 67, "extra_field_not_allowed": true}
+{"field_68": 68, "extra_field_not_allowed": true}
+{"field_69": 69, "extra_field_not_allowed": true}
+{"field_70": 70, "extra_field_not_allowed": true}
+{"field_71": 71, "extra_field_not_allowed": true}
+{"field_72": 72, "extra_field_not_allowed": true}
+{"field_73": 73, "extra_field_not_allowed": true}
+{"field_74": 74, "extra_field_not_allowed": true}
+{"field_75": 75, "extra_field_not_allowed": true}
+{"field_76": 76, "extra_field_not_allowed": true}
+{"field_77": 77, "extra_field_not_allowed": true}
+{"field_78": 78, "extra_field_not_allowed": true}
+{"field_79": 79, "extra_field_not_allowed": true}
+{"field_80": 80, "extra_field_not_allowed": true}
+{"field_81": 81, "extra_field_not_allowed": true}
+{"field_82": 82, "extra_field_not_allowed": true}
+{"field_83": 83, "extra_field_not_allowed": true}
+{"field_84": 84, "extra_field_not_allowed": true}
+{"field_85": 85, "extra_field_not_allowed": true}
+{"field_86": 86, "extra_field_not_allowed": true}
+{"field_87": 87, "extra_field_not_allowed": true}
+{"field_88": 88, "extra_field_not_allowed": true}
+{"field_89": 89, "extra_field_not_allowed": true}
+{"field_90": 90, "extra_field_not_allowed": true}
+{"field_91": 91, "extra_field_not_allowed": true}
+{"field_92": 92, "extra_field_not_allowed": true}
+{"field_93": 93, "extra_field_not_allowed": true}
+{"field_94": 94, "extra_field_not_allowed": true}
+{"field_95": 95, "extra_field_not_allowed": true}
+{"field_96": 96, "extra_field_not_allowed": true}
+{"field_97": 97, "extra_field_not_allowed": true}
+{"field_98": 98, "extra_field_not_allowed": true}
+{"field_99": 99, "extra_field_not_allowed": true}
+{"field_100": 100, "extra_field_not_allowed": true}
+{"field_101": 101, "extra_field_not_allowed": true}
+{"field_102": 102, "extra_field_not_allowed": true}
+{"field_103": 103, "extra_field_not_allowed": true}
+{"field_104": 104, "extra_field_not_allowed": true}
+{"field_105": 105, "extra_field_not_allowed": true}
+{"field_106": 106, "extra_field_not_allowed": true}
+{"field_107": 107, "extra_field_not_allowed": true}
+{"field_108": 108, "extra_field_not_allowed": true}
+{"field_109": 109, "extra_field_not_allowed": true}
+{"field_110": 110, "extra_field_not_allowed": true}
+{"field_111": 111, "extra_field_not_allowed": true}
+{"field_112": 112, "extra_field_not_allowed": true}
+{"field_113": 113, "extra_field_not_allowed": true}
+{"field_114": 114, "extra_field_not_allowed": true}
+{"field_115": 115, "extra_field_not_allowed": true}
+{"field_116": 116, "extra_field_not_allowed": true}
+{"field_117": 117, "extra_field_not_allowed": true}
+{"field_118": 118, "extra_field_not_allowed": true}
+{"field_119": 119, "extra_field_not_allowed": true}
+{"field_120": 120, "extra_field_not_allowed": true}
+{"field_121": 121, "extra_field_not_allowed": true}
+{"field_122": 122, "extra_field_not_allowed": true}
+{"field_123": 123, "extra_field_not_allowed": true}
+{"field_124": 124, "extra_field_not_allowed": true}
+{"field_125": 125, "extra_field_not_allowed": true}
+{"field_126": 126, "extra_field_not_allowed": true}
+{"field_127": 127, "extra_field_not_allowed": true}
+{"field_128": 128, "extra_field_not_allowed": true}
+{"field_129": 129, "extra_field_not_allowed": true}
+{"field_130": 130, "extra_field_not_allowed": true}
+{"field_131": 131, "extra_field_not_allowed": true}
+{"field_132": 132, "extra_field_not_allowed": true}
+{"field_133": 133, "extra_field_not_allowed": true}
+{"field_134": 134, "extra_field_not_allowed": true}
+{"field_135": 135, "extra_field_not_allowed": true}
+{"field_136": 136, "extra_field_not_allowed": true}
+{"field_137": 137, "extra_field_not_allowed": true}
+{"field_138": 138, "extra_field_not_allowed": true}
+{"field_139": 139, "extra_field_not_allowed": true}
+{"field_140": 140, "extra_field_not_allowed": true}
+{"field_141": 141, "extra_field_not_allowed": true}
+{"field_142": 142, "extra_field_not_allowed": true}
+{"field_143": 143, "extra_field_not_allowed": true}
+{"field_144": 144, "extra_field_not_allowed": true}
+{"field_145": 145, "extra_field_not_allowed": true}
+{"field_146": 146, "extra_field_not_allowed": true}
+{"field_147": 147, "extra_field_not_allowed": true}
+{"field_148": 148, "extra_field_not_allowed": true}
+{"field_149": 149, "extra_field_not_allowed": true}
+{"field_150": 150, "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/wrong_value.ndjson
new file mode 100644
index 000000000..b5793f0ab
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-large/wrong_value.ndjson
@@ -0,0 +1,150 @@
+{"id": 99999, "field_1": 99999}
+{"id": 99999, "field_2": 99999}
+{"id": 99999, "field_3": 99999}
+{"id": 99999, "field_4": 99999}
+{"id": 99999, "field_5": 99999}
+{"id": 99999, "field_6": 99999}
+{"id": 99999, "field_7": 99999}
+{"id": 99999, "field_8": 99999}
+{"id": 99999, "field_9": 99999}
+{"id": 99999, "field_10": 99999}
+{"id": 99999, "field_11": 99999}
+{"id": 99999, "field_12": 99999}
+{"id": 99999, "field_13": 99999}
+{"id": 99999, "field_14": 99999}
+{"id": 99999, "field_15": 99999}
+{"id": 99999, "field_16": 99999}
+{"id": 99999, "field_17": 99999}
+{"id": 99999, "field_18": 99999}
+{"id": 99999, "field_19": 99999}
+{"id": 99999, "field_20": 99999}
+{"id": 99999, "field_21": 99999}
+{"id": 99999, "field_22": 99999}
+{"id": 99999, "field_23": 99999}
+{"id": 99999, "field_24": 99999}
+{"id": 99999, "field_25": 99999}
+{"id": 99999, "field_26": 99999}
+{"id": 99999, "field_27": 99999}
+{"id": 99999, "field_28": 99999}
+{"id": 99999, "field_29": 99999}
+{"id": 99999, "field_30": 99999}
+{"id": 99999, "field_31": 99999}
+{"id": 99999, "field_32": 99999}
+{"id": 99999, "field_33": 99999}
+{"id": 99999, "field_34": 99999}
+{"id": 99999, "field_35": 99999}
+{"id": 99999, "field_36": 99999}
+{"id": 99999, "field_37": 99999}
+{"id": 99999, "field_38": 99999}
+{"id": 99999, "field_39": 99999}
+{"id": 99999, "field_40": 99999}
+{"id": 99999, "field_41": 99999}
+{"id": 99999, "field_42": 99999}
+{"id": 99999, "field_43": 99999}
+{"id": 99999, "field_44": 99999}
+{"id": 99999, "field_45": 99999}
+{"id": 99999, "field_46": 99999}
+{"id": 99999, "field_47": 99999}
+{"id": 99999, "field_48": 99999}
+{"id": 99999, "field_49": 99999}
+{"id": 99999, "field_50": 99999}
+{"id": 99999, "field_51": 99999}
+{"id": 99999, "field_52": 99999}
+{"id": 99999, "field_53": 99999}
+{"id": 99999, "field_54": 99999}
+{"id": 99999, "field_55": 99999}
+{"id": 99999, "field_56": 99999}
+{"id": 99999, "field_57": 99999}
+{"id": 99999, "field_58": 99999}
+{"id": 99999, "field_59": 99999}
+{"id": 99999, "field_60": 99999}
+{"id": 99999, "field_61": 99999}
+{"id": 99999, "field_62": 99999}
+{"id": 99999, "field_63": 99999}
+{"id": 99999, "field_64": 99999}
+{"id": 99999, "field_65": 99999}
+{"id": 99999, "field_66": 99999}
+{"id": 99999, "field_67": 99999}
+{"id": 99999, "field_68": 99999}
+{"id": 99999, "field_69": 99999}
+{"id": 99999, "field_70": 99999}
+{"id": 99999, "field_71": 99999}
+{"id": 99999, "field_72": 99999}
+{"id": 99999, "field_73": 99999}
+{"id": 99999, "field_74": 99999}
+{"id": 99999, "field_75": 99999}
+{"id": 99999, "field_76": 99999}
+{"id": 99999, "field_77": 99999}
+{"id": 99999, "field_78": 99999}
+{"id": 99999, "field_79": 99999}
+{"id": 99999, "field_80": 99999}
+{"id": 99999, "field_81": 99999}
+{"id": 99999, "field_82": 99999}
+{"id": 99999, "field_83": 99999}
+{"id": 99999, "field_84": 99999}
+{"id": 99999, "field_85": 99999}
+{"id": 99999, "field_86": 99999}
+{"id": 99999, "field_87": 99999}
+{"id": 99999, "field_88": 99999}
+{"id": 99999, "field_89": 99999}
+{"id": 99999, "field_90": 99999}
+{"id": 99999, "field_91": 99999}
+{"id": 99999, "field_92": 99999}
+{"id": 99999, "field_93": 99999}
+{"id": 99999, "field_94": 99999}
+{"id": 99999, "field_95": 99999}
+{"id": 99999, "field_96": 99999}
+{"id": 99999, "field_97": 99999}
+{"id": 99999, "field_98": 99999}
+{"id": 99999, "field_99": 99999}
+{"id": 99999, "field_100": 99999}
+{"id": 99999, "field_101": 99999}
+{"id": 99999, "field_102": 99999}
+{"id": 99999, "field_103": 99999}
+{"id": 99999, "field_104": 99999}
+{"id": 99999, "field_105": 99999}
+{"id": 99999, "field_106": 99999}
+{"id": 99999, "field_107": 99999}
+{"id": 99999, "field_108": 99999}
+{"id": 99999, "field_109": 99999}
+{"id": 99999, "field_110": 99999}
+{"id": 99999, "field_111": 99999}
+{"id": 99999, "field_112": 99999}
+{"id": 99999, "field_113": 99999}
+{"id": 99999, "field_114": 99999}
+{"id": 99999, "field_115": 99999}
+{"id": 99999, "field_116": 99999}
+{"id": 99999, "field_117": 99999}
+{"id": 99999, "field_118": 99999}
+{"id": 99999, "field_119": 99999}
+{"id": 99999, "field_120": 99999}
+{"id": 99999, "field_121": 99999}
+{"id": 99999, "field_122": 99999}
+{"id": 99999, "field_123": 99999}
+{"id": 99999, "field_124": 99999}
+{"id": 99999, "field_125": 99999}
+{"id": 99999, "field_126": 99999}
+{"id": 99999, "field_127": 99999}
+{"id": 99999, "field_128": 99999}
+{"id": 99999, "field_129": 99999}
+{"id": 99999, "field_130": 99999}
+{"id": 99999, "field_131": 99999}
+{"id": 99999, "field_132": 99999}
+{"id": 99999, "field_133": 99999}
+{"id": 99999, "field_134": 99999}
+{"id": 99999, "field_135": 99999}
+{"id": 99999, "field_136": 99999}
+{"id": 99999, "field_137": 99999}
+{"id": 99999, "field_138": 99999}
+{"id": 99999, "field_139": 99999}
+{"id": 99999, "field_140": 99999}
+{"id": 99999, "field_141": 99999}
+{"id": 99999, "field_142": 99999}
+{"id": 99999, "field_143": 99999}
+{"id": 99999, "field_144": 99999}
+{"id": 99999, "field_145": 99999}
+{"id": 99999, "field_146": 99999}
+{"id": 99999, "field_147": 99999}
+{"id": 99999, "field_148": 99999}
+{"id": 99999, "field_149": 99999}
+{"id": 99999, "field_150": 99999}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/canonical.ndjson
new file mode 100644
index 000000000..49833919e
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/canonical.ndjson
@@ -0,0 +1,15 @@
+{"id": 1, "field_1": 1}
+{"id": 2, "field_2": 2}
+{"id": 3, "field_3": 3}
+{"id": 4, "field_4": 4}
+{"id": 5, "field_5": 5}
+{"id": 6, "field_6": 6}
+{"id": 7, "field_7": 7}
+{"id": 8, "field_8": 8}
+{"id": 9, "field_9": 9}
+{"id": 10, "field_10": 10}
+{"id": 11, "field_11": 11}
+{"id": 12, "field_12": 12}
+{"id": 13, "field_13": 13}
+{"id": 14, "field_14": 14}
+{"id": 15, "field_15": 15}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/duplicate.ndjson
new file mode 100644
index 000000000..e93b72b31
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/duplicate.ndjson
@@ -0,0 +1,16 @@
+{"id": 1, "field_1": 1}
+{"id": 2, "field_2": 2}
+{"id": 3, "field_3": 3}
+{"id": 4, "field_4": 4}
+{"id": 5, "field_5": 5}
+{"id": 6, "field_6": 6}
+{"id": 7, "field_7": 7}
+{"id": 8, "field_8": 8}
+{"id": 9, "field_9": 9}
+{"id": 10, "field_10": 10}
+{"id": 11, "field_11": 11}
+{"id": 12, "field_12": 12}
+{"id": 13, "field_13": 13}
+{"id": 14, "field_14": 14}
+{"id": 15, "field_15": 15}
+{"id": 15, "field_15": 15}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/extra.ndjson
new file mode 100644
index 000000000..d30636b21
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/extra.ndjson
@@ -0,0 +1,16 @@
+{"id": 1, "field_1": 1}
+{"id": 2, "field_2": 2}
+{"id": 3, "field_3": 3}
+{"id": 4, "field_4": 4}
+{"id": 5, "field_5": 5}
+{"id": 6, "field_6": 6}
+{"id": 7, "field_7": 7}
+{"id": 8, "field_8": 8}
+{"id": 9, "field_9": 9}
+{"id": 10, "field_10": 10}
+{"id": 11, "field_11": 11}
+{"id": 12, "field_12": 12}
+{"id": 13, "field_13": 13}
+{"id": 14, "field_14": 14}
+{"id": 15, "field_15": 15}
+{"id": 1, "field_1": 1, "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/family_mut.ndjson
new file mode 100644
index 000000000..fa9adf15a
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/family_mut.ndjson
@@ -0,0 +1,15 @@
+{"id": 1, "field_1": 1}
+{"id": 2, "field_2": 2}
+{"id": 3, "field_3": 3}
+{"id": 4, "field_4": 4}
+{"id": 5, "field_5": 5}
+{"id": 6, "field_6": 6}
+{"id": 7, "field_7": 7}
+{"id": 8, "field_8": 8}
+{"id": 9, "field_9": 9}
+{"id": 10, "field_10": 10}
+{"id": 11, "field_11": 11}
+{"id": 12, "field_12": 12}
+{"id": 13, "field_13": 13}
+{"id": 14, "field_14": 14}
+{"id": 15, "field_1": 1}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/missing.ndjson
new file mode 100644
index 000000000..9c336cfe4
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/missing.ndjson
@@ -0,0 +1,14 @@
+{"id": 1, "field_1": 1}
+{"id": 2, "field_2": 2}
+{"id": 3, "field_3": 3}
+{"id": 4, "field_4": 4}
+{"id": 5, "field_5": 5}
+{"id": 6, "field_6": 6}
+{"id": 7, "field_7": 7}
+{"id": 8, "field_8": 8}
+{"id": 9, "field_9": 9}
+{"id": 10, "field_10": 10}
+{"id": 11, "field_11": 11}
+{"id": 12, "field_12": 12}
+{"id": 13, "field_13": 13}
+{"id": 14, "field_14": 14}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/wrong_field_set.ndjson
new file mode 100644
index 000000000..68878e296
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/wrong_field_set.ndjson
@@ -0,0 +1,15 @@
+{"field_1": 1, "extra_field_not_allowed": true}
+{"field_2": 2, "extra_field_not_allowed": true}
+{"field_3": 3, "extra_field_not_allowed": true}
+{"field_4": 4, "extra_field_not_allowed": true}
+{"field_5": 5, "extra_field_not_allowed": true}
+{"field_6": 6, "extra_field_not_allowed": true}
+{"field_7": 7, "extra_field_not_allowed": true}
+{"field_8": 8, "extra_field_not_allowed": true}
+{"field_9": 9, "extra_field_not_allowed": true}
+{"field_10": 10, "extra_field_not_allowed": true}
+{"field_11": 11, "extra_field_not_allowed": true}
+{"field_12": 12, "extra_field_not_allowed": true}
+{"field_13": 13, "extra_field_not_allowed": true}
+{"field_14": 14, "extra_field_not_allowed": true}
+{"field_15": 15, "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/wrong_value.ndjson
new file mode 100644
index 000000000..17b3dabe7
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/irregular-medium/wrong_value.ndjson
@@ -0,0 +1,15 @@
+{"id": 99999, "field_1": 99999}
+{"id": 99999, "field_2": 99999}
+{"id": 99999, "field_3": 99999}
+{"id": 99999, "field_4": 99999}
+{"id": 99999, "field_5": 99999}
+{"id": 99999, "field_6": 99999}
+{"id": 99999, "field_7": 99999}
+{"id": 99999, "field_8": 99999}
+{"id": 99999, "field_9": 99999}
+{"id": 99999, "field_10": 99999}
+{"id": 99999, "field_11": 99999}
+{"id": 99999, "field_12": 99999}
+{"id": 99999, "field_13": 99999}
+{"id": 99999, "field_14": 99999}
+{"id": 99999, "field_15": 99999}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/canonical.ndjson
new file mode 100644
index 000000000..2979a1d60
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/canonical.ndjson
@@ -0,0 +1,48 @@
+{"role": "guest", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/login", "status": "deny"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/duplicate.ndjson
new file mode 100644
index 000000000..e0822852b
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/duplicate.ndjson
@@ -0,0 +1,49 @@
+{"role": "guest", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/login", "status": "deny"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/extra.ndjson
new file mode 100644
index 000000000..91fee43fb
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/extra.ndjson
@@ -0,0 +1,49 @@
+{"role": "guest", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "GET", "endpoint": "/api", "status": "deny", "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/family_mut.ndjson
new file mode 100644
index 000000000..29dca6369
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/family_mut.ndjson
@@ -0,0 +1,47 @@
+{"role": "guest", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/login", "status": "deny"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/missing.ndjson
new file mode 100644
index 000000000..d9e446458
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/missing.ndjson
@@ -0,0 +1,47 @@
+{"role": "guest", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "guest", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "member", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/web", "status": "deny"}
+{"role": "admin", "method": "DELETE", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "GET", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "POST", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/web", "status": "deny"}
+{"role": "superuser", "method": "PUT", "endpoint": "/login", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/api", "status": "deny"}
+{"role": "superuser", "method": "DELETE", "endpoint": "/web", "status": "deny"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/wrong_field_set.ndjson
new file mode 100644
index 000000000..70a18c745
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/wrong_field_set.ndjson
@@ -0,0 +1,48 @@
+{"role": "guest", "method": "GET", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "GET", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "GET", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "POST", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "POST", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "POST", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "PUT", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "PUT", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "PUT", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "DELETE", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "DELETE", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "DELETE", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "GET", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "GET", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "GET", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "POST", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "POST", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "POST", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "PUT", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "PUT", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "PUT", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "DELETE", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "DELETE", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "DELETE", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "GET", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "GET", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "GET", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "POST", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "POST", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "POST", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "PUT", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "PUT", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "PUT", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "DELETE", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "DELETE", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "DELETE", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "GET", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "GET", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "GET", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "POST", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "POST", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "POST", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "PUT", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "PUT", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "PUT", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "DELETE", "endpoint": "/api", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "DELETE", "endpoint": "/web", "status": "deny", "extra_field_not_allowed": true}
+{"role": "superuser", "method": "DELETE", "endpoint": "/login", "status": "deny", "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/wrong_value.ndjson
new file mode 100644
index 000000000..943b6ef54
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-large/wrong_value.ndjson
@@ -0,0 +1,48 @@
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "endpoint": "WRONG", "status": "WRONG"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/canonical.ndjson
new file mode 100644
index 000000000..ab4a58a7f
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/canonical.ndjson
@@ -0,0 +1,6 @@
+{"role": "guest", "method": "GET", "status": "deny"}
+{"role": "guest", "method": "POST", "status": "deny"}
+{"role": "member", "method": "GET", "status": "deny"}
+{"role": "member", "method": "POST", "status": "deny"}
+{"role": "admin", "method": "GET", "status": "deny"}
+{"role": "admin", "method": "POST", "status": "deny"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/duplicate.ndjson
new file mode 100644
index 000000000..c7bfee26b
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/duplicate.ndjson
@@ -0,0 +1,7 @@
+{"role": "guest", "method": "GET", "status": "deny"}
+{"role": "guest", "method": "POST", "status": "deny"}
+{"role": "member", "method": "GET", "status": "deny"}
+{"role": "member", "method": "POST", "status": "deny"}
+{"role": "admin", "method": "GET", "status": "deny"}
+{"role": "admin", "method": "POST", "status": "deny"}
+{"role": "admin", "method": "POST", "status": "deny"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/extra.ndjson
new file mode 100644
index 000000000..e3d157df8
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/extra.ndjson
@@ -0,0 +1,7 @@
+{"role": "guest", "method": "GET", "status": "deny"}
+{"role": "guest", "method": "POST", "status": "deny"}
+{"role": "member", "method": "GET", "status": "deny"}
+{"role": "member", "method": "POST", "status": "deny"}
+{"role": "admin", "method": "GET", "status": "deny"}
+{"role": "admin", "method": "POST", "status": "deny"}
+{"role": "guest", "method": "GET", "status": "deny", "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/family_mut.ndjson
new file mode 100644
index 000000000..d411d929a
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/family_mut.ndjson
@@ -0,0 +1,5 @@
+{"role": "guest", "method": "POST", "status": "deny"}
+{"role": "member", "method": "GET", "status": "deny"}
+{"role": "member", "method": "POST", "status": "deny"}
+{"role": "admin", "method": "GET", "status": "deny"}
+{"role": "admin", "method": "POST", "status": "deny"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/missing.ndjson
new file mode 100644
index 000000000..833841e2f
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/missing.ndjson
@@ -0,0 +1,5 @@
+{"role": "guest", "method": "GET", "status": "deny"}
+{"role": "guest", "method": "POST", "status": "deny"}
+{"role": "member", "method": "GET", "status": "deny"}
+{"role": "member", "method": "POST", "status": "deny"}
+{"role": "admin", "method": "GET", "status": "deny"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/wrong_field_set.ndjson
new file mode 100644
index 000000000..4293923d4
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/wrong_field_set.ndjson
@@ -0,0 +1,6 @@
+{"role": "guest", "method": "GET", "status": "deny", "extra_field_not_allowed": true}
+{"role": "guest", "method": "POST", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "GET", "status": "deny", "extra_field_not_allowed": true}
+{"role": "member", "method": "POST", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "GET", "status": "deny", "extra_field_not_allowed": true}
+{"role": "admin", "method": "POST", "status": "deny", "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/wrong_value.ndjson
new file mode 100644
index 000000000..a90dc4374
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/matrix-medium/wrong_value.ndjson
@@ -0,0 +1,6 @@
+{"role": "WRONG", "method": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "status": "WRONG"}
+{"role": "WRONG", "method": "WRONG", "status": "WRONG"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/canonical.ndjson
new file mode 100644
index 000000000..f65aa3120
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/canonical.ndjson
@@ -0,0 +1,250 @@
+{"id": 1, "type": "std"}
+{"id": 2, "type": "std"}
+{"id": 3, "type": "std"}
+{"id": 4, "type": "std"}
+{"id": 5, "type": "std"}
+{"id": 6, "type": "std"}
+{"id": 7, "type": "std"}
+{"id": 8, "type": "std"}
+{"id": 9, "type": "std"}
+{"id": 10, "type": "std"}
+{"id": 11, "type": "std"}
+{"id": 12, "type": "std"}
+{"id": 13, "type": "std"}
+{"id": 14, "type": "std"}
+{"id": 15, "type": "std"}
+{"id": 16, "type": "std"}
+{"id": 17, "type": "std"}
+{"id": 18, "type": "std"}
+{"id": 19, "type": "std"}
+{"id": 20, "type": "std"}
+{"id": 21, "type": "std"}
+{"id": 22, "type": "std"}
+{"id": 23, "type": "std"}
+{"id": 24, "type": "std"}
+{"id": 25, "type": "std"}
+{"id": 26, "type": "std"}
+{"id": 27, "type": "std"}
+{"id": 28, "type": "std"}
+{"id": 29, "type": "std"}
+{"id": 30, "type": "std"}
+{"id": 31, "type": "std"}
+{"id": 32, "type": "std"}
+{"id": 33, "type": "std"}
+{"id": 34, "type": "std"}
+{"id": 35, "type": "std"}
+{"id": 36, "type": "std"}
+{"id": 37, "type": "std"}
+{"id": 38, "type": "std"}
+{"id": 39, "type": "std"}
+{"id": 40, "type": "std"}
+{"id": 41, "type": "std"}
+{"id": 42, "type": "std"}
+{"id": 43, "type": "std"}
+{"id": 44, "type": "std"}
+{"id": 45, "type": "std"}
+{"id": 46, "type": "std"}
+{"id": 47, "type": "std"}
+{"id": 48, "type": "std"}
+{"id": 49, "type": "std"}
+{"id": 50, "type": "std"}
+{"id": 51, "type": "std"}
+{"id": 52, "type": "std"}
+{"id": 53, "type": "std"}
+{"id": 54, "type": "std"}
+{"id": 55, "type": "std"}
+{"id": 56, "type": "std"}
+{"id": 57, "type": "std"}
+{"id": 58, "type": "std"}
+{"id": 59, "type": "std"}
+{"id": 60, "type": "std"}
+{"id": 61, "type": "std"}
+{"id": 62, "type": "std"}
+{"id": 63, "type": "std"}
+{"id": 64, "type": "std"}
+{"id": 65, "type": "std"}
+{"id": 66, "type": "std"}
+{"id": 67, "type": "std"}
+{"id": 68, "type": "std"}
+{"id": 69, "type": "std"}
+{"id": 70, "type": "std"}
+{"id": 71, "type": "std"}
+{"id": 72, "type": "std"}
+{"id": 73, "type": "std"}
+{"id": 74, "type": "std"}
+{"id": 75, "type": "std"}
+{"id": 76, "type": "std"}
+{"id": 77, "type": "std"}
+{"id": 78, "type": "std"}
+{"id": 79, "type": "std"}
+{"id": 80, "type": "std"}
+{"id": 81, "type": "std"}
+{"id": 82, "type": "std"}
+{"id": 83, "type": "std"}
+{"id": 84, "type": "std"}
+{"id": 85, "type": "std"}
+{"id": 86, "type": "std"}
+{"id": 87, "type": "std"}
+{"id": 88, "type": "std"}
+{"id": 89, "type": "std"}
+{"id": 90, "type": "std"}
+{"id": 91, "type": "std"}
+{"id": 92, "type": "std"}
+{"id": 93, "type": "std"}
+{"id": 94, "type": "std"}
+{"id": 95, "type": "std"}
+{"id": 96, "type": "std"}
+{"id": 97, "type": "std"}
+{"id": 98, "type": "std"}
+{"id": 99, "type": "std"}
+{"id": 100, "type": "std"}
+{"id": 101, "type": "std"}
+{"id": 102, "type": "std"}
+{"id": 103, "type": "std"}
+{"id": 104, "type": "std"}
+{"id": 105, "type": "std"}
+{"id": 106, "type": "std"}
+{"id": 107, "type": "std"}
+{"id": 108, "type": "std"}
+{"id": 109, "type": "std"}
+{"id": 110, "type": "std"}
+{"id": 111, "type": "std"}
+{"id": 112, "type": "std"}
+{"id": 113, "type": "std"}
+{"id": 114, "type": "std"}
+{"id": 115, "type": "std"}
+{"id": 116, "type": "std"}
+{"id": 117, "type": "std"}
+{"id": 118, "type": "std"}
+{"id": 119, "type": "std"}
+{"id": 120, "type": "std"}
+{"id": 121, "type": "std"}
+{"id": 122, "type": "std"}
+{"id": 123, "type": "std"}
+{"id": 124, "type": "std"}
+{"id": 125, "type": "std"}
+{"id": 126, "type": "std"}
+{"id": 127, "type": "std"}
+{"id": 128, "type": "std"}
+{"id": 129, "type": "std"}
+{"id": 130, "type": "std"}
+{"id": 131, "type": "std"}
+{"id": 132, "type": "std"}
+{"id": 133, "type": "std"}
+{"id": 134, "type": "std"}
+{"id": 135, "type": "std"}
+{"id": 136, "type": "std"}
+{"id": 137, "type": "std"}
+{"id": 138, "type": "std"}
+{"id": 139, "type": "std"}
+{"id": 140, "type": "std"}
+{"id": 141, "type": "std"}
+{"id": 142, "type": "std"}
+{"id": 143, "type": "std"}
+{"id": 144, "type": "std"}
+{"id": 145, "type": "std"}
+{"id": 146, "type": "std"}
+{"id": 147, "type": "std"}
+{"id": 148, "type": "std"}
+{"id": 149, "type": "std"}
+{"id": 150, "type": "std"}
+{"id": 151, "type": "std"}
+{"id": 152, "type": "std"}
+{"id": 153, "type": "std"}
+{"id": 154, "type": "std"}
+{"id": 155, "type": "std"}
+{"id": 156, "type": "std"}
+{"id": 157, "type": "std"}
+{"id": 158, "type": "std"}
+{"id": 159, "type": "std"}
+{"id": 160, "type": "std"}
+{"id": 161, "type": "std"}
+{"id": 162, "type": "std"}
+{"id": 163, "type": "std"}
+{"id": 164, "type": "std"}
+{"id": 165, "type": "std"}
+{"id": 166, "type": "std"}
+{"id": 167, "type": "std"}
+{"id": 168, "type": "std"}
+{"id": 169, "type": "std"}
+{"id": 170, "type": "std"}
+{"id": 171, "type": "std"}
+{"id": 172, "type": "std"}
+{"id": 173, "type": "std"}
+{"id": 174, "type": "std"}
+{"id": 175, "type": "std"}
+{"id": 176, "type": "std"}
+{"id": 177, "type": "std"}
+{"id": 178, "type": "std"}
+{"id": 179, "type": "std"}
+{"id": 180, "type": "std"}
+{"id": 181, "type": "std"}
+{"id": 182, "type": "std"}
+{"id": 183, "type": "std"}
+{"id": 184, "type": "std"}
+{"id": 185, "type": "std"}
+{"id": 186, "type": "std"}
+{"id": 187, "type": "std"}
+{"id": 188, "type": "std"}
+{"id": 189, "type": "std"}
+{"id": 190, "type": "std"}
+{"id": 191, "type": "std"}
+{"id": 192, "type": "std"}
+{"id": 193, "type": "std"}
+{"id": 194, "type": "std"}
+{"id": 195, "type": "std"}
+{"id": 196, "type": "std"}
+{"id": 197, "type": "std"}
+{"id": 198, "type": "std"}
+{"id": 199, "type": "std"}
+{"id": 200, "type": "std"}
+{"id": 201, "type": "exc", "error_code": 500}
+{"id": 202, "type": "exc", "error_code": 500}
+{"id": 203, "type": "exc", "error_code": 500}
+{"id": 204, "type": "exc", "error_code": 500}
+{"id": 205, "type": "exc", "error_code": 500}
+{"id": 206, "type": "exc", "error_code": 500}
+{"id": 207, "type": "exc", "error_code": 500}
+{"id": 208, "type": "exc", "error_code": 500}
+{"id": 209, "type": "exc", "error_code": 500}
+{"id": 210, "type": "exc", "error_code": 500}
+{"id": 211, "type": "exc", "error_code": 500}
+{"id": 212, "type": "exc", "error_code": 500}
+{"id": 213, "type": "exc", "error_code": 500}
+{"id": 214, "type": "exc", "error_code": 500}
+{"id": 215, "type": "exc", "error_code": 500}
+{"id": 216, "type": "exc", "error_code": 500}
+{"id": 217, "type": "exc", "error_code": 500}
+{"id": 218, "type": "exc", "error_code": 500}
+{"id": 219, "type": "exc", "error_code": 500}
+{"id": 220, "type": "exc", "error_code": 500}
+{"id": 221, "type": "exc", "error_code": 500}
+{"id": 222, "type": "exc", "error_code": 500}
+{"id": 223, "type": "exc", "error_code": 500}
+{"id": 224, "type": "exc", "error_code": 500}
+{"id": 225, "type": "exc", "error_code": 500}
+{"id": 226, "type": "exc", "error_code": 500}
+{"id": 227, "type": "exc", "error_code": 500}
+{"id": 228, "type": "exc", "error_code": 500}
+{"id": 229, "type": "exc", "error_code": 500}
+{"id": 230, "type": "exc", "error_code": 500}
+{"id": 231, "type": "exc", "error_code": 500}
+{"id": 232, "type": "exc", "error_code": 500}
+{"id": 233, "type": "exc", "error_code": 500}
+{"id": 234, "type": "exc", "error_code": 500}
+{"id": 235, "type": "exc", "error_code": 500}
+{"id": 236, "type": "exc", "error_code": 500}
+{"id": 237, "type": "exc", "error_code": 500}
+{"id": 238, "type": "exc", "error_code": 500}
+{"id": 239, "type": "exc", "error_code": 500}
+{"id": 240, "type": "exc", "error_code": 500}
+{"id": 241, "type": "exc", "error_code": 500}
+{"id": 242, "type": "exc", "error_code": 500}
+{"id": 243, "type": "exc", "error_code": 500}
+{"id": 244, "type": "exc", "error_code": 500}
+{"id": 245, "type": "exc", "error_code": 500}
+{"id": 246, "type": "exc", "error_code": 500}
+{"id": 247, "type": "exc", "error_code": 500}
+{"id": 248, "type": "exc", "error_code": 500}
+{"id": 249, "type": "exc", "error_code": 500}
+{"id": 250, "type": "exc", "error_code": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/duplicate.ndjson
new file mode 100644
index 000000000..3cec9c677
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/duplicate.ndjson
@@ -0,0 +1,251 @@
+{"id": 1, "type": "std"}
+{"id": 2, "type": "std"}
+{"id": 3, "type": "std"}
+{"id": 4, "type": "std"}
+{"id": 5, "type": "std"}
+{"id": 6, "type": "std"}
+{"id": 7, "type": "std"}
+{"id": 8, "type": "std"}
+{"id": 9, "type": "std"}
+{"id": 10, "type": "std"}
+{"id": 11, "type": "std"}
+{"id": 12, "type": "std"}
+{"id": 13, "type": "std"}
+{"id": 14, "type": "std"}
+{"id": 15, "type": "std"}
+{"id": 16, "type": "std"}
+{"id": 17, "type": "std"}
+{"id": 18, "type": "std"}
+{"id": 19, "type": "std"}
+{"id": 20, "type": "std"}
+{"id": 21, "type": "std"}
+{"id": 22, "type": "std"}
+{"id": 23, "type": "std"}
+{"id": 24, "type": "std"}
+{"id": 25, "type": "std"}
+{"id": 26, "type": "std"}
+{"id": 27, "type": "std"}
+{"id": 28, "type": "std"}
+{"id": 29, "type": "std"}
+{"id": 30, "type": "std"}
+{"id": 31, "type": "std"}
+{"id": 32, "type": "std"}
+{"id": 33, "type": "std"}
+{"id": 34, "type": "std"}
+{"id": 35, "type": "std"}
+{"id": 36, "type": "std"}
+{"id": 37, "type": "std"}
+{"id": 38, "type": "std"}
+{"id": 39, "type": "std"}
+{"id": 40, "type": "std"}
+{"id": 41, "type": "std"}
+{"id": 42, "type": "std"}
+{"id": 43, "type": "std"}
+{"id": 44, "type": "std"}
+{"id": 45, "type": "std"}
+{"id": 46, "type": "std"}
+{"id": 47, "type": "std"}
+{"id": 48, "type": "std"}
+{"id": 49, "type": "std"}
+{"id": 50, "type": "std"}
+{"id": 51, "type": "std"}
+{"id": 52, "type": "std"}
+{"id": 53, "type": "std"}
+{"id": 54, "type": "std"}
+{"id": 55, "type": "std"}
+{"id": 56, "type": "std"}
+{"id": 57, "type": "std"}
+{"id": 58, "type": "std"}
+{"id": 59, "type": "std"}
+{"id": 60, "type": "std"}
+{"id": 61, "type": "std"}
+{"id": 62, "type": "std"}
+{"id": 63, "type": "std"}
+{"id": 64, "type": "std"}
+{"id": 65, "type": "std"}
+{"id": 66, "type": "std"}
+{"id": 67, "type": "std"}
+{"id": 68, "type": "std"}
+{"id": 69, "type": "std"}
+{"id": 70, "type": "std"}
+{"id": 71, "type": "std"}
+{"id": 72, "type": "std"}
+{"id": 73, "type": "std"}
+{"id": 74, "type": "std"}
+{"id": 75, "type": "std"}
+{"id": 76, "type": "std"}
+{"id": 77, "type": "std"}
+{"id": 78, "type": "std"}
+{"id": 79, "type": "std"}
+{"id": 80, "type": "std"}
+{"id": 81, "type": "std"}
+{"id": 82, "type": "std"}
+{"id": 83, "type": "std"}
+{"id": 84, "type": "std"}
+{"id": 85, "type": "std"}
+{"id": 86, "type": "std"}
+{"id": 87, "type": "std"}
+{"id": 88, "type": "std"}
+{"id": 89, "type": "std"}
+{"id": 90, "type": "std"}
+{"id": 91, "type": "std"}
+{"id": 92, "type": "std"}
+{"id": 93, "type": "std"}
+{"id": 94, "type": "std"}
+{"id": 95, "type": "std"}
+{"id": 96, "type": "std"}
+{"id": 97, "type": "std"}
+{"id": 98, "type": "std"}
+{"id": 99, "type": "std"}
+{"id": 100, "type": "std"}
+{"id": 101, "type": "std"}
+{"id": 102, "type": "std"}
+{"id": 103, "type": "std"}
+{"id": 104, "type": "std"}
+{"id": 105, "type": "std"}
+{"id": 106, "type": "std"}
+{"id": 107, "type": "std"}
+{"id": 108, "type": "std"}
+{"id": 109, "type": "std"}
+{"id": 110, "type": "std"}
+{"id": 111, "type": "std"}
+{"id": 112, "type": "std"}
+{"id": 113, "type": "std"}
+{"id": 114, "type": "std"}
+{"id": 115, "type": "std"}
+{"id": 116, "type": "std"}
+{"id": 117, "type": "std"}
+{"id": 118, "type": "std"}
+{"id": 119, "type": "std"}
+{"id": 120, "type": "std"}
+{"id": 121, "type": "std"}
+{"id": 122, "type": "std"}
+{"id": 123, "type": "std"}
+{"id": 124, "type": "std"}
+{"id": 125, "type": "std"}
+{"id": 126, "type": "std"}
+{"id": 127, "type": "std"}
+{"id": 128, "type": "std"}
+{"id": 129, "type": "std"}
+{"id": 130, "type": "std"}
+{"id": 131, "type": "std"}
+{"id": 132, "type": "std"}
+{"id": 133, "type": "std"}
+{"id": 134, "type": "std"}
+{"id": 135, "type": "std"}
+{"id": 136, "type": "std"}
+{"id": 137, "type": "std"}
+{"id": 138, "type": "std"}
+{"id": 139, "type": "std"}
+{"id": 140, "type": "std"}
+{"id": 141, "type": "std"}
+{"id": 142, "type": "std"}
+{"id": 143, "type": "std"}
+{"id": 144, "type": "std"}
+{"id": 145, "type": "std"}
+{"id": 146, "type": "std"}
+{"id": 147, "type": "std"}
+{"id": 148, "type": "std"}
+{"id": 149, "type": "std"}
+{"id": 150, "type": "std"}
+{"id": 151, "type": "std"}
+{"id": 152, "type": "std"}
+{"id": 153, "type": "std"}
+{"id": 154, "type": "std"}
+{"id": 155, "type": "std"}
+{"id": 156, "type": "std"}
+{"id": 157, "type": "std"}
+{"id": 158, "type": "std"}
+{"id": 159, "type": "std"}
+{"id": 160, "type": "std"}
+{"id": 161, "type": "std"}
+{"id": 162, "type": "std"}
+{"id": 163, "type": "std"}
+{"id": 164, "type": "std"}
+{"id": 165, "type": "std"}
+{"id": 166, "type": "std"}
+{"id": 167, "type": "std"}
+{"id": 168, "type": "std"}
+{"id": 169, "type": "std"}
+{"id": 170, "type": "std"}
+{"id": 171, "type": "std"}
+{"id": 172, "type": "std"}
+{"id": 173, "type": "std"}
+{"id": 174, "type": "std"}
+{"id": 175, "type": "std"}
+{"id": 176, "type": "std"}
+{"id": 177, "type": "std"}
+{"id": 178, "type": "std"}
+{"id": 179, "type": "std"}
+{"id": 180, "type": "std"}
+{"id": 181, "type": "std"}
+{"id": 182, "type": "std"}
+{"id": 183, "type": "std"}
+{"id": 184, "type": "std"}
+{"id": 185, "type": "std"}
+{"id": 186, "type": "std"}
+{"id": 187, "type": "std"}
+{"id": 188, "type": "std"}
+{"id": 189, "type": "std"}
+{"id": 190, "type": "std"}
+{"id": 191, "type": "std"}
+{"id": 192, "type": "std"}
+{"id": 193, "type": "std"}
+{"id": 194, "type": "std"}
+{"id": 195, "type": "std"}
+{"id": 196, "type": "std"}
+{"id": 197, "type": "std"}
+{"id": 198, "type": "std"}
+{"id": 199, "type": "std"}
+{"id": 200, "type": "std"}
+{"id": 201, "type": "exc", "error_code": 500}
+{"id": 202, "type": "exc", "error_code": 500}
+{"id": 203, "type": "exc", "error_code": 500}
+{"id": 204, "type": "exc", "error_code": 500}
+{"id": 205, "type": "exc", "error_code": 500}
+{"id": 206, "type": "exc", "error_code": 500}
+{"id": 207, "type": "exc", "error_code": 500}
+{"id": 208, "type": "exc", "error_code": 500}
+{"id": 209, "type": "exc", "error_code": 500}
+{"id": 210, "type": "exc", "error_code": 500}
+{"id": 211, "type": "exc", "error_code": 500}
+{"id": 212, "type": "exc", "error_code": 500}
+{"id": 213, "type": "exc", "error_code": 500}
+{"id": 214, "type": "exc", "error_code": 500}
+{"id": 215, "type": "exc", "error_code": 500}
+{"id": 216, "type": "exc", "error_code": 500}
+{"id": 217, "type": "exc", "error_code": 500}
+{"id": 218, "type": "exc", "error_code": 500}
+{"id": 219, "type": "exc", "error_code": 500}
+{"id": 220, "type": "exc", "error_code": 500}
+{"id": 221, "type": "exc", "error_code": 500}
+{"id": 222, "type": "exc", "error_code": 500}
+{"id": 223, "type": "exc", "error_code": 500}
+{"id": 224, "type": "exc", "error_code": 500}
+{"id": 225, "type": "exc", "error_code": 500}
+{"id": 226, "type": "exc", "error_code": 500}
+{"id": 227, "type": "exc", "error_code": 500}
+{"id": 228, "type": "exc", "error_code": 500}
+{"id": 229, "type": "exc", "error_code": 500}
+{"id": 230, "type": "exc", "error_code": 500}
+{"id": 231, "type": "exc", "error_code": 500}
+{"id": 232, "type": "exc", "error_code": 500}
+{"id": 233, "type": "exc", "error_code": 500}
+{"id": 234, "type": "exc", "error_code": 500}
+{"id": 235, "type": "exc", "error_code": 500}
+{"id": 236, "type": "exc", "error_code": 500}
+{"id": 237, "type": "exc", "error_code": 500}
+{"id": 238, "type": "exc", "error_code": 500}
+{"id": 239, "type": "exc", "error_code": 500}
+{"id": 240, "type": "exc", "error_code": 500}
+{"id": 241, "type": "exc", "error_code": 500}
+{"id": 242, "type": "exc", "error_code": 500}
+{"id": 243, "type": "exc", "error_code": 500}
+{"id": 244, "type": "exc", "error_code": 500}
+{"id": 245, "type": "exc", "error_code": 500}
+{"id": 246, "type": "exc", "error_code": 500}
+{"id": 247, "type": "exc", "error_code": 500}
+{"id": 248, "type": "exc", "error_code": 500}
+{"id": 249, "type": "exc", "error_code": 500}
+{"id": 250, "type": "exc", "error_code": 500}
+{"id": 250, "type": "exc", "error_code": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/extra.ndjson
new file mode 100644
index 000000000..8468809a4
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/extra.ndjson
@@ -0,0 +1,251 @@
+{"id": 1, "type": "std"}
+{"id": 2, "type": "std"}
+{"id": 3, "type": "std"}
+{"id": 4, "type": "std"}
+{"id": 5, "type": "std"}
+{"id": 6, "type": "std"}
+{"id": 7, "type": "std"}
+{"id": 8, "type": "std"}
+{"id": 9, "type": "std"}
+{"id": 10, "type": "std"}
+{"id": 11, "type": "std"}
+{"id": 12, "type": "std"}
+{"id": 13, "type": "std"}
+{"id": 14, "type": "std"}
+{"id": 15, "type": "std"}
+{"id": 16, "type": "std"}
+{"id": 17, "type": "std"}
+{"id": 18, "type": "std"}
+{"id": 19, "type": "std"}
+{"id": 20, "type": "std"}
+{"id": 21, "type": "std"}
+{"id": 22, "type": "std"}
+{"id": 23, "type": "std"}
+{"id": 24, "type": "std"}
+{"id": 25, "type": "std"}
+{"id": 26, "type": "std"}
+{"id": 27, "type": "std"}
+{"id": 28, "type": "std"}
+{"id": 29, "type": "std"}
+{"id": 30, "type": "std"}
+{"id": 31, "type": "std"}
+{"id": 32, "type": "std"}
+{"id": 33, "type": "std"}
+{"id": 34, "type": "std"}
+{"id": 35, "type": "std"}
+{"id": 36, "type": "std"}
+{"id": 37, "type": "std"}
+{"id": 38, "type": "std"}
+{"id": 39, "type": "std"}
+{"id": 40, "type": "std"}
+{"id": 41, "type": "std"}
+{"id": 42, "type": "std"}
+{"id": 43, "type": "std"}
+{"id": 44, "type": "std"}
+{"id": 45, "type": "std"}
+{"id": 46, "type": "std"}
+{"id": 47, "type": "std"}
+{"id": 48, "type": "std"}
+{"id": 49, "type": "std"}
+{"id": 50, "type": "std"}
+{"id": 51, "type": "std"}
+{"id": 52, "type": "std"}
+{"id": 53, "type": "std"}
+{"id": 54, "type": "std"}
+{"id": 55, "type": "std"}
+{"id": 56, "type": "std"}
+{"id": 57, "type": "std"}
+{"id": 58, "type": "std"}
+{"id": 59, "type": "std"}
+{"id": 60, "type": "std"}
+{"id": 61, "type": "std"}
+{"id": 62, "type": "std"}
+{"id": 63, "type": "std"}
+{"id": 64, "type": "std"}
+{"id": 65, "type": "std"}
+{"id": 66, "type": "std"}
+{"id": 67, "type": "std"}
+{"id": 68, "type": "std"}
+{"id": 69, "type": "std"}
+{"id": 70, "type": "std"}
+{"id": 71, "type": "std"}
+{"id": 72, "type": "std"}
+{"id": 73, "type": "std"}
+{"id": 74, "type": "std"}
+{"id": 75, "type": "std"}
+{"id": 76, "type": "std"}
+{"id": 77, "type": "std"}
+{"id": 78, "type": "std"}
+{"id": 79, "type": "std"}
+{"id": 80, "type": "std"}
+{"id": 81, "type": "std"}
+{"id": 82, "type": "std"}
+{"id": 83, "type": "std"}
+{"id": 84, "type": "std"}
+{"id": 85, "type": "std"}
+{"id": 86, "type": "std"}
+{"id": 87, "type": "std"}
+{"id": 88, "type": "std"}
+{"id": 89, "type": "std"}
+{"id": 90, "type": "std"}
+{"id": 91, "type": "std"}
+{"id": 92, "type": "std"}
+{"id": 93, "type": "std"}
+{"id": 94, "type": "std"}
+{"id": 95, "type": "std"}
+{"id": 96, "type": "std"}
+{"id": 97, "type": "std"}
+{"id": 98, "type": "std"}
+{"id": 99, "type": "std"}
+{"id": 100, "type": "std"}
+{"id": 101, "type": "std"}
+{"id": 102, "type": "std"}
+{"id": 103, "type": "std"}
+{"id": 104, "type": "std"}
+{"id": 105, "type": "std"}
+{"id": 106, "type": "std"}
+{"id": 107, "type": "std"}
+{"id": 108, "type": "std"}
+{"id": 109, "type": "std"}
+{"id": 110, "type": "std"}
+{"id": 111, "type": "std"}
+{"id": 112, "type": "std"}
+{"id": 113, "type": "std"}
+{"id": 114, "type": "std"}
+{"id": 115, "type": "std"}
+{"id": 116, "type": "std"}
+{"id": 117, "type": "std"}
+{"id": 118, "type": "std"}
+{"id": 119, "type": "std"}
+{"id": 120, "type": "std"}
+{"id": 121, "type": "std"}
+{"id": 122, "type": "std"}
+{"id": 123, "type": "std"}
+{"id": 124, "type": "std"}
+{"id": 125, "type": "std"}
+{"id": 126, "type": "std"}
+{"id": 127, "type": "std"}
+{"id": 128, "type": "std"}
+{"id": 129, "type": "std"}
+{"id": 130, "type": "std"}
+{"id": 131, "type": "std"}
+{"id": 132, "type": "std"}
+{"id": 133, "type": "std"}
+{"id": 134, "type": "std"}
+{"id": 135, "type": "std"}
+{"id": 136, "type": "std"}
+{"id": 137, "type": "std"}
+{"id": 138, "type": "std"}
+{"id": 139, "type": "std"}
+{"id": 140, "type": "std"}
+{"id": 141, "type": "std"}
+{"id": 142, "type": "std"}
+{"id": 143, "type": "std"}
+{"id": 144, "type": "std"}
+{"id": 145, "type": "std"}
+{"id": 146, "type": "std"}
+{"id": 147, "type": "std"}
+{"id": 148, "type": "std"}
+{"id": 149, "type": "std"}
+{"id": 150, "type": "std"}
+{"id": 151, "type": "std"}
+{"id": 152, "type": "std"}
+{"id": 153, "type": "std"}
+{"id": 154, "type": "std"}
+{"id": 155, "type": "std"}
+{"id": 156, "type": "std"}
+{"id": 157, "type": "std"}
+{"id": 158, "type": "std"}
+{"id": 159, "type": "std"}
+{"id": 160, "type": "std"}
+{"id": 161, "type": "std"}
+{"id": 162, "type": "std"}
+{"id": 163, "type": "std"}
+{"id": 164, "type": "std"}
+{"id": 165, "type": "std"}
+{"id": 166, "type": "std"}
+{"id": 167, "type": "std"}
+{"id": 168, "type": "std"}
+{"id": 169, "type": "std"}
+{"id": 170, "type": "std"}
+{"id": 171, "type": "std"}
+{"id": 172, "type": "std"}
+{"id": 173, "type": "std"}
+{"id": 174, "type": "std"}
+{"id": 175, "type": "std"}
+{"id": 176, "type": "std"}
+{"id": 177, "type": "std"}
+{"id": 178, "type": "std"}
+{"id": 179, "type": "std"}
+{"id": 180, "type": "std"}
+{"id": 181, "type": "std"}
+{"id": 182, "type": "std"}
+{"id": 183, "type": "std"}
+{"id": 184, "type": "std"}
+{"id": 185, "type": "std"}
+{"id": 186, "type": "std"}
+{"id": 187, "type": "std"}
+{"id": 188, "type": "std"}
+{"id": 189, "type": "std"}
+{"id": 190, "type": "std"}
+{"id": 191, "type": "std"}
+{"id": 192, "type": "std"}
+{"id": 193, "type": "std"}
+{"id": 194, "type": "std"}
+{"id": 195, "type": "std"}
+{"id": 196, "type": "std"}
+{"id": 197, "type": "std"}
+{"id": 198, "type": "std"}
+{"id": 199, "type": "std"}
+{"id": 200, "type": "std"}
+{"id": 201, "type": "exc", "error_code": 500}
+{"id": 202, "type": "exc", "error_code": 500}
+{"id": 203, "type": "exc", "error_code": 500}
+{"id": 204, "type": "exc", "error_code": 500}
+{"id": 205, "type": "exc", "error_code": 500}
+{"id": 206, "type": "exc", "error_code": 500}
+{"id": 207, "type": "exc", "error_code": 500}
+{"id": 208, "type": "exc", "error_code": 500}
+{"id": 209, "type": "exc", "error_code": 500}
+{"id": 210, "type": "exc", "error_code": 500}
+{"id": 211, "type": "exc", "error_code": 500}
+{"id": 212, "type": "exc", "error_code": 500}
+{"id": 213, "type": "exc", "error_code": 500}
+{"id": 214, "type": "exc", "error_code": 500}
+{"id": 215, "type": "exc", "error_code": 500}
+{"id": 216, "type": "exc", "error_code": 500}
+{"id": 217, "type": "exc", "error_code": 500}
+{"id": 218, "type": "exc", "error_code": 500}
+{"id": 219, "type": "exc", "error_code": 500}
+{"id": 220, "type": "exc", "error_code": 500}
+{"id": 221, "type": "exc", "error_code": 500}
+{"id": 222, "type": "exc", "error_code": 500}
+{"id": 223, "type": "exc", "error_code": 500}
+{"id": 224, "type": "exc", "error_code": 500}
+{"id": 225, "type": "exc", "error_code": 500}
+{"id": 226, "type": "exc", "error_code": 500}
+{"id": 227, "type": "exc", "error_code": 500}
+{"id": 228, "type": "exc", "error_code": 500}
+{"id": 229, "type": "exc", "error_code": 500}
+{"id": 230, "type": "exc", "error_code": 500}
+{"id": 231, "type": "exc", "error_code": 500}
+{"id": 232, "type": "exc", "error_code": 500}
+{"id": 233, "type": "exc", "error_code": 500}
+{"id": 234, "type": "exc", "error_code": 500}
+{"id": 235, "type": "exc", "error_code": 500}
+{"id": 236, "type": "exc", "error_code": 500}
+{"id": 237, "type": "exc", "error_code": 500}
+{"id": 238, "type": "exc", "error_code": 500}
+{"id": 239, "type": "exc", "error_code": 500}
+{"id": 240, "type": "exc", "error_code": 500}
+{"id": 241, "type": "exc", "error_code": 500}
+{"id": 242, "type": "exc", "error_code": 500}
+{"id": 243, "type": "exc", "error_code": 500}
+{"id": 244, "type": "exc", "error_code": 500}
+{"id": 245, "type": "exc", "error_code": 500}
+{"id": 246, "type": "exc", "error_code": 500}
+{"id": 247, "type": "exc", "error_code": 500}
+{"id": 248, "type": "exc", "error_code": 500}
+{"id": 249, "type": "exc", "error_code": 500}
+{"id": 250, "type": "exc", "error_code": 500}
+{"id": 1, "type": "std", "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/family_mut.ndjson
new file mode 100644
index 000000000..78b89c143
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/family_mut.ndjson
@@ -0,0 +1,249 @@
+{"id": 1, "type": "std"}
+{"id": 2, "type": "std"}
+{"id": 3, "type": "std"}
+{"id": 4, "type": "std"}
+{"id": 5, "type": "std"}
+{"id": 6, "type": "std"}
+{"id": 7, "type": "std"}
+{"id": 8, "type": "std"}
+{"id": 9, "type": "std"}
+{"id": 10, "type": "std"}
+{"id": 11, "type": "std"}
+{"id": 12, "type": "std"}
+{"id": 13, "type": "std"}
+{"id": 14, "type": "std"}
+{"id": 15, "type": "std"}
+{"id": 16, "type": "std"}
+{"id": 17, "type": "std"}
+{"id": 18, "type": "std"}
+{"id": 19, "type": "std"}
+{"id": 20, "type": "std"}
+{"id": 21, "type": "std"}
+{"id": 22, "type": "std"}
+{"id": 23, "type": "std"}
+{"id": 24, "type": "std"}
+{"id": 25, "type": "std"}
+{"id": 26, "type": "std"}
+{"id": 27, "type": "std"}
+{"id": 28, "type": "std"}
+{"id": 29, "type": "std"}
+{"id": 30, "type": "std"}
+{"id": 31, "type": "std"}
+{"id": 32, "type": "std"}
+{"id": 33, "type": "std"}
+{"id": 34, "type": "std"}
+{"id": 35, "type": "std"}
+{"id": 36, "type": "std"}
+{"id": 37, "type": "std"}
+{"id": 38, "type": "std"}
+{"id": 39, "type": "std"}
+{"id": 40, "type": "std"}
+{"id": 41, "type": "std"}
+{"id": 42, "type": "std"}
+{"id": 43, "type": "std"}
+{"id": 44, "type": "std"}
+{"id": 45, "type": "std"}
+{"id": 46, "type": "std"}
+{"id": 47, "type": "std"}
+{"id": 48, "type": "std"}
+{"id": 49, "type": "std"}
+{"id": 50, "type": "std"}
+{"id": 51, "type": "std"}
+{"id": 52, "type": "std"}
+{"id": 53, "type": "std"}
+{"id": 54, "type": "std"}
+{"id": 55, "type": "std"}
+{"id": 56, "type": "std"}
+{"id": 57, "type": "std"}
+{"id": 58, "type": "std"}
+{"id": 59, "type": "std"}
+{"id": 60, "type": "std"}
+{"id": 61, "type": "std"}
+{"id": 62, "type": "std"}
+{"id": 63, "type": "std"}
+{"id": 64, "type": "std"}
+{"id": 65, "type": "std"}
+{"id": 66, "type": "std"}
+{"id": 67, "type": "std"}
+{"id": 68, "type": "std"}
+{"id": 69, "type": "std"}
+{"id": 70, "type": "std"}
+{"id": 71, "type": "std"}
+{"id": 72, "type": "std"}
+{"id": 73, "type": "std"}
+{"id": 74, "type": "std"}
+{"id": 75, "type": "std"}
+{"id": 76, "type": "std"}
+{"id": 77, "type": "std"}
+{"id": 78, "type": "std"}
+{"id": 79, "type": "std"}
+{"id": 80, "type": "std"}
+{"id": 81, "type": "std"}
+{"id": 82, "type": "std"}
+{"id": 83, "type": "std"}
+{"id": 84, "type": "std"}
+{"id": 85, "type": "std"}
+{"id": 86, "type": "std"}
+{"id": 87, "type": "std"}
+{"id": 88, "type": "std"}
+{"id": 89, "type": "std"}
+{"id": 90, "type": "std"}
+{"id": 91, "type": "std"}
+{"id": 92, "type": "std"}
+{"id": 93, "type": "std"}
+{"id": 94, "type": "std"}
+{"id": 95, "type": "std"}
+{"id": 96, "type": "std"}
+{"id": 97, "type": "std"}
+{"id": 98, "type": "std"}
+{"id": 99, "type": "std"}
+{"id": 100, "type": "std"}
+{"id": 101, "type": "std"}
+{"id": 102, "type": "std"}
+{"id": 103, "type": "std"}
+{"id": 104, "type": "std"}
+{"id": 105, "type": "std"}
+{"id": 106, "type": "std"}
+{"id": 107, "type": "std"}
+{"id": 108, "type": "std"}
+{"id": 109, "type": "std"}
+{"id": 110, "type": "std"}
+{"id": 111, "type": "std"}
+{"id": 112, "type": "std"}
+{"id": 113, "type": "std"}
+{"id": 114, "type": "std"}
+{"id": 115, "type": "std"}
+{"id": 116, "type": "std"}
+{"id": 117, "type": "std"}
+{"id": 118, "type": "std"}
+{"id": 119, "type": "std"}
+{"id": 120, "type": "std"}
+{"id": 121, "type": "std"}
+{"id": 122, "type": "std"}
+{"id": 123, "type": "std"}
+{"id": 124, "type": "std"}
+{"id": 125, "type": "std"}
+{"id": 126, "type": "std"}
+{"id": 127, "type": "std"}
+{"id": 128, "type": "std"}
+{"id": 129, "type": "std"}
+{"id": 130, "type": "std"}
+{"id": 131, "type": "std"}
+{"id": 132, "type": "std"}
+{"id": 133, "type": "std"}
+{"id": 134, "type": "std"}
+{"id": 135, "type": "std"}
+{"id": 136, "type": "std"}
+{"id": 137, "type": "std"}
+{"id": 138, "type": "std"}
+{"id": 139, "type": "std"}
+{"id": 140, "type": "std"}
+{"id": 141, "type": "std"}
+{"id": 142, "type": "std"}
+{"id": 143, "type": "std"}
+{"id": 144, "type": "std"}
+{"id": 145, "type": "std"}
+{"id": 146, "type": "std"}
+{"id": 147, "type": "std"}
+{"id": 148, "type": "std"}
+{"id": 149, "type": "std"}
+{"id": 150, "type": "std"}
+{"id": 151, "type": "std"}
+{"id": 152, "type": "std"}
+{"id": 153, "type": "std"}
+{"id": 154, "type": "std"}
+{"id": 155, "type": "std"}
+{"id": 156, "type": "std"}
+{"id": 157, "type": "std"}
+{"id": 158, "type": "std"}
+{"id": 159, "type": "std"}
+{"id": 160, "type": "std"}
+{"id": 161, "type": "std"}
+{"id": 162, "type": "std"}
+{"id": 163, "type": "std"}
+{"id": 164, "type": "std"}
+{"id": 165, "type": "std"}
+{"id": 166, "type": "std"}
+{"id": 167, "type": "std"}
+{"id": 168, "type": "std"}
+{"id": 169, "type": "std"}
+{"id": 170, "type": "std"}
+{"id": 171, "type": "std"}
+{"id": 172, "type": "std"}
+{"id": 173, "type": "std"}
+{"id": 174, "type": "std"}
+{"id": 175, "type": "std"}
+{"id": 176, "type": "std"}
+{"id": 177, "type": "std"}
+{"id": 178, "type": "std"}
+{"id": 179, "type": "std"}
+{"id": 180, "type": "std"}
+{"id": 181, "type": "std"}
+{"id": 182, "type": "std"}
+{"id": 183, "type": "std"}
+{"id": 184, "type": "std"}
+{"id": 185, "type": "std"}
+{"id": 186, "type": "std"}
+{"id": 187, "type": "std"}
+{"id": 188, "type": "std"}
+{"id": 189, "type": "std"}
+{"id": 190, "type": "std"}
+{"id": 191, "type": "std"}
+{"id": 192, "type": "std"}
+{"id": 193, "type": "std"}
+{"id": 194, "type": "std"}
+{"id": 195, "type": "std"}
+{"id": 196, "type": "std"}
+{"id": 197, "type": "std"}
+{"id": 198, "type": "std"}
+{"id": 199, "type": "std"}
+{"id": 200, "type": "std"}
+{"id": 201, "type": "exc", "error_code": 500}
+{"id": 202, "type": "exc", "error_code": 500}
+{"id": 203, "type": "exc", "error_code": 500}
+{"id": 204, "type": "exc", "error_code": 500}
+{"id": 205, "type": "exc", "error_code": 500}
+{"id": 206, "type": "exc", "error_code": 500}
+{"id": 207, "type": "exc", "error_code": 500}
+{"id": 208, "type": "exc", "error_code": 500}
+{"id": 209, "type": "exc", "error_code": 500}
+{"id": 210, "type": "exc", "error_code": 500}
+{"id": 211, "type": "exc", "error_code": 500}
+{"id": 212, "type": "exc", "error_code": 500}
+{"id": 213, "type": "exc", "error_code": 500}
+{"id": 214, "type": "exc", "error_code": 500}
+{"id": 215, "type": "exc", "error_code": 500}
+{"id": 216, "type": "exc", "error_code": 500}
+{"id": 217, "type": "exc", "error_code": 500}
+{"id": 218, "type": "exc", "error_code": 500}
+{"id": 219, "type": "exc", "error_code": 500}
+{"id": 220, "type": "exc", "error_code": 500}
+{"id": 221, "type": "exc", "error_code": 500}
+{"id": 222, "type": "exc", "error_code": 500}
+{"id": 223, "type": "exc", "error_code": 500}
+{"id": 224, "type": "exc", "error_code": 500}
+{"id": 225, "type": "exc", "error_code": 500}
+{"id": 226, "type": "exc", "error_code": 500}
+{"id": 227, "type": "exc", "error_code": 500}
+{"id": 228, "type": "exc", "error_code": 500}
+{"id": 229, "type": "exc", "error_code": 500}
+{"id": 230, "type": "exc", "error_code": 500}
+{"id": 231, "type": "exc", "error_code": 500}
+{"id": 232, "type": "exc", "error_code": 500}
+{"id": 233, "type": "exc", "error_code": 500}
+{"id": 234, "type": "exc", "error_code": 500}
+{"id": 235, "type": "exc", "error_code": 500}
+{"id": 236, "type": "exc", "error_code": 500}
+{"id": 237, "type": "exc", "error_code": 500}
+{"id": 238, "type": "exc", "error_code": 500}
+{"id": 239, "type": "exc", "error_code": 500}
+{"id": 240, "type": "exc", "error_code": 500}
+{"id": 241, "type": "exc", "error_code": 500}
+{"id": 242, "type": "exc", "error_code": 500}
+{"id": 243, "type": "exc", "error_code": 500}
+{"id": 244, "type": "exc", "error_code": 500}
+{"id": 245, "type": "exc", "error_code": 500}
+{"id": 246, "type": "exc", "error_code": 500}
+{"id": 247, "type": "exc", "error_code": 500}
+{"id": 248, "type": "exc", "error_code": 500}
+{"id": 249, "type": "exc", "error_code": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/missing.ndjson
new file mode 100644
index 000000000..78b89c143
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/missing.ndjson
@@ -0,0 +1,249 @@
+{"id": 1, "type": "std"}
+{"id": 2, "type": "std"}
+{"id": 3, "type": "std"}
+{"id": 4, "type": "std"}
+{"id": 5, "type": "std"}
+{"id": 6, "type": "std"}
+{"id": 7, "type": "std"}
+{"id": 8, "type": "std"}
+{"id": 9, "type": "std"}
+{"id": 10, "type": "std"}
+{"id": 11, "type": "std"}
+{"id": 12, "type": "std"}
+{"id": 13, "type": "std"}
+{"id": 14, "type": "std"}
+{"id": 15, "type": "std"}
+{"id": 16, "type": "std"}
+{"id": 17, "type": "std"}
+{"id": 18, "type": "std"}
+{"id": 19, "type": "std"}
+{"id": 20, "type": "std"}
+{"id": 21, "type": "std"}
+{"id": 22, "type": "std"}
+{"id": 23, "type": "std"}
+{"id": 24, "type": "std"}
+{"id": 25, "type": "std"}
+{"id": 26, "type": "std"}
+{"id": 27, "type": "std"}
+{"id": 28, "type": "std"}
+{"id": 29, "type": "std"}
+{"id": 30, "type": "std"}
+{"id": 31, "type": "std"}
+{"id": 32, "type": "std"}
+{"id": 33, "type": "std"}
+{"id": 34, "type": "std"}
+{"id": 35, "type": "std"}
+{"id": 36, "type": "std"}
+{"id": 37, "type": "std"}
+{"id": 38, "type": "std"}
+{"id": 39, "type": "std"}
+{"id": 40, "type": "std"}
+{"id": 41, "type": "std"}
+{"id": 42, "type": "std"}
+{"id": 43, "type": "std"}
+{"id": 44, "type": "std"}
+{"id": 45, "type": "std"}
+{"id": 46, "type": "std"}
+{"id": 47, "type": "std"}
+{"id": 48, "type": "std"}
+{"id": 49, "type": "std"}
+{"id": 50, "type": "std"}
+{"id": 51, "type": "std"}
+{"id": 52, "type": "std"}
+{"id": 53, "type": "std"}
+{"id": 54, "type": "std"}
+{"id": 55, "type": "std"}
+{"id": 56, "type": "std"}
+{"id": 57, "type": "std"}
+{"id": 58, "type": "std"}
+{"id": 59, "type": "std"}
+{"id": 60, "type": "std"}
+{"id": 61, "type": "std"}
+{"id": 62, "type": "std"}
+{"id": 63, "type": "std"}
+{"id": 64, "type": "std"}
+{"id": 65, "type": "std"}
+{"id": 66, "type": "std"}
+{"id": 67, "type": "std"}
+{"id": 68, "type": "std"}
+{"id": 69, "type": "std"}
+{"id": 70, "type": "std"}
+{"id": 71, "type": "std"}
+{"id": 72, "type": "std"}
+{"id": 73, "type": "std"}
+{"id": 74, "type": "std"}
+{"id": 75, "type": "std"}
+{"id": 76, "type": "std"}
+{"id": 77, "type": "std"}
+{"id": 78, "type": "std"}
+{"id": 79, "type": "std"}
+{"id": 80, "type": "std"}
+{"id": 81, "type": "std"}
+{"id": 82, "type": "std"}
+{"id": 83, "type": "std"}
+{"id": 84, "type": "std"}
+{"id": 85, "type": "std"}
+{"id": 86, "type": "std"}
+{"id": 87, "type": "std"}
+{"id": 88, "type": "std"}
+{"id": 89, "type": "std"}
+{"id": 90, "type": "std"}
+{"id": 91, "type": "std"}
+{"id": 92, "type": "std"}
+{"id": 93, "type": "std"}
+{"id": 94, "type": "std"}
+{"id": 95, "type": "std"}
+{"id": 96, "type": "std"}
+{"id": 97, "type": "std"}
+{"id": 98, "type": "std"}
+{"id": 99, "type": "std"}
+{"id": 100, "type": "std"}
+{"id": 101, "type": "std"}
+{"id": 102, "type": "std"}
+{"id": 103, "type": "std"}
+{"id": 104, "type": "std"}
+{"id": 105, "type": "std"}
+{"id": 106, "type": "std"}
+{"id": 107, "type": "std"}
+{"id": 108, "type": "std"}
+{"id": 109, "type": "std"}
+{"id": 110, "type": "std"}
+{"id": 111, "type": "std"}
+{"id": 112, "type": "std"}
+{"id": 113, "type": "std"}
+{"id": 114, "type": "std"}
+{"id": 115, "type": "std"}
+{"id": 116, "type": "std"}
+{"id": 117, "type": "std"}
+{"id": 118, "type": "std"}
+{"id": 119, "type": "std"}
+{"id": 120, "type": "std"}
+{"id": 121, "type": "std"}
+{"id": 122, "type": "std"}
+{"id": 123, "type": "std"}
+{"id": 124, "type": "std"}
+{"id": 125, "type": "std"}
+{"id": 126, "type": "std"}
+{"id": 127, "type": "std"}
+{"id": 128, "type": "std"}
+{"id": 129, "type": "std"}
+{"id": 130, "type": "std"}
+{"id": 131, "type": "std"}
+{"id": 132, "type": "std"}
+{"id": 133, "type": "std"}
+{"id": 134, "type": "std"}
+{"id": 135, "type": "std"}
+{"id": 136, "type": "std"}
+{"id": 137, "type": "std"}
+{"id": 138, "type": "std"}
+{"id": 139, "type": "std"}
+{"id": 140, "type": "std"}
+{"id": 141, "type": "std"}
+{"id": 142, "type": "std"}
+{"id": 143, "type": "std"}
+{"id": 144, "type": "std"}
+{"id": 145, "type": "std"}
+{"id": 146, "type": "std"}
+{"id": 147, "type": "std"}
+{"id": 148, "type": "std"}
+{"id": 149, "type": "std"}
+{"id": 150, "type": "std"}
+{"id": 151, "type": "std"}
+{"id": 152, "type": "std"}
+{"id": 153, "type": "std"}
+{"id": 154, "type": "std"}
+{"id": 155, "type": "std"}
+{"id": 156, "type": "std"}
+{"id": 157, "type": "std"}
+{"id": 158, "type": "std"}
+{"id": 159, "type": "std"}
+{"id": 160, "type": "std"}
+{"id": 161, "type": "std"}
+{"id": 162, "type": "std"}
+{"id": 163, "type": "std"}
+{"id": 164, "type": "std"}
+{"id": 165, "type": "std"}
+{"id": 166, "type": "std"}
+{"id": 167, "type": "std"}
+{"id": 168, "type": "std"}
+{"id": 169, "type": "std"}
+{"id": 170, "type": "std"}
+{"id": 171, "type": "std"}
+{"id": 172, "type": "std"}
+{"id": 173, "type": "std"}
+{"id": 174, "type": "std"}
+{"id": 175, "type": "std"}
+{"id": 176, "type": "std"}
+{"id": 177, "type": "std"}
+{"id": 178, "type": "std"}
+{"id": 179, "type": "std"}
+{"id": 180, "type": "std"}
+{"id": 181, "type": "std"}
+{"id": 182, "type": "std"}
+{"id": 183, "type": "std"}
+{"id": 184, "type": "std"}
+{"id": 185, "type": "std"}
+{"id": 186, "type": "std"}
+{"id": 187, "type": "std"}
+{"id": 188, "type": "std"}
+{"id": 189, "type": "std"}
+{"id": 190, "type": "std"}
+{"id": 191, "type": "std"}
+{"id": 192, "type": "std"}
+{"id": 193, "type": "std"}
+{"id": 194, "type": "std"}
+{"id": 195, "type": "std"}
+{"id": 196, "type": "std"}
+{"id": 197, "type": "std"}
+{"id": 198, "type": "std"}
+{"id": 199, "type": "std"}
+{"id": 200, "type": "std"}
+{"id": 201, "type": "exc", "error_code": 500}
+{"id": 202, "type": "exc", "error_code": 500}
+{"id": 203, "type": "exc", "error_code": 500}
+{"id": 204, "type": "exc", "error_code": 500}
+{"id": 205, "type": "exc", "error_code": 500}
+{"id": 206, "type": "exc", "error_code": 500}
+{"id": 207, "type": "exc", "error_code": 500}
+{"id": 208, "type": "exc", "error_code": 500}
+{"id": 209, "type": "exc", "error_code": 500}
+{"id": 210, "type": "exc", "error_code": 500}
+{"id": 211, "type": "exc", "error_code": 500}
+{"id": 212, "type": "exc", "error_code": 500}
+{"id": 213, "type": "exc", "error_code": 500}
+{"id": 214, "type": "exc", "error_code": 500}
+{"id": 215, "type": "exc", "error_code": 500}
+{"id": 216, "type": "exc", "error_code": 500}
+{"id": 217, "type": "exc", "error_code": 500}
+{"id": 218, "type": "exc", "error_code": 500}
+{"id": 219, "type": "exc", "error_code": 500}
+{"id": 220, "type": "exc", "error_code": 500}
+{"id": 221, "type": "exc", "error_code": 500}
+{"id": 222, "type": "exc", "error_code": 500}
+{"id": 223, "type": "exc", "error_code": 500}
+{"id": 224, "type": "exc", "error_code": 500}
+{"id": 225, "type": "exc", "error_code": 500}
+{"id": 226, "type": "exc", "error_code": 500}
+{"id": 227, "type": "exc", "error_code": 500}
+{"id": 228, "type": "exc", "error_code": 500}
+{"id": 229, "type": "exc", "error_code": 500}
+{"id": 230, "type": "exc", "error_code": 500}
+{"id": 231, "type": "exc", "error_code": 500}
+{"id": 232, "type": "exc", "error_code": 500}
+{"id": 233, "type": "exc", "error_code": 500}
+{"id": 234, "type": "exc", "error_code": 500}
+{"id": 235, "type": "exc", "error_code": 500}
+{"id": 236, "type": "exc", "error_code": 500}
+{"id": 237, "type": "exc", "error_code": 500}
+{"id": 238, "type": "exc", "error_code": 500}
+{"id": 239, "type": "exc", "error_code": 500}
+{"id": 240, "type": "exc", "error_code": 500}
+{"id": 241, "type": "exc", "error_code": 500}
+{"id": 242, "type": "exc", "error_code": 500}
+{"id": 243, "type": "exc", "error_code": 500}
+{"id": 244, "type": "exc", "error_code": 500}
+{"id": 245, "type": "exc", "error_code": 500}
+{"id": 246, "type": "exc", "error_code": 500}
+{"id": 247, "type": "exc", "error_code": 500}
+{"id": 248, "type": "exc", "error_code": 500}
+{"id": 249, "type": "exc", "error_code": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/wrong_field_set.ndjson
new file mode 100644
index 000000000..43a399e87
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/wrong_field_set.ndjson
@@ -0,0 +1,250 @@
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/wrong_value.ndjson
new file mode 100644
index 000000000..84ae8fadc
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-large/wrong_value.ndjson
@@ -0,0 +1,250 @@
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/canonical.ndjson
new file mode 100644
index 000000000..bf6f1ad63
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/canonical.ndjson
@@ -0,0 +1,25 @@
+{"id": 1, "type": "std"}
+{"id": 2, "type": "std"}
+{"id": 3, "type": "std"}
+{"id": 4, "type": "std"}
+{"id": 5, "type": "std"}
+{"id": 6, "type": "std"}
+{"id": 7, "type": "std"}
+{"id": 8, "type": "std"}
+{"id": 9, "type": "std"}
+{"id": 10, "type": "std"}
+{"id": 11, "type": "std"}
+{"id": 12, "type": "std"}
+{"id": 13, "type": "std"}
+{"id": 14, "type": "std"}
+{"id": 15, "type": "std"}
+{"id": 16, "type": "std"}
+{"id": 17, "type": "std"}
+{"id": 18, "type": "std"}
+{"id": 19, "type": "std"}
+{"id": 20, "type": "std"}
+{"id": 21, "type": "exc", "error_code": 500}
+{"id": 22, "type": "exc", "error_code": 500}
+{"id": 23, "type": "exc", "error_code": 500}
+{"id": 24, "type": "exc", "error_code": 500}
+{"id": 25, "type": "exc", "error_code": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/duplicate.ndjson
new file mode 100644
index 000000000..a30f648c9
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/duplicate.ndjson
@@ -0,0 +1,26 @@
+{"id": 1, "type": "std"}
+{"id": 2, "type": "std"}
+{"id": 3, "type": "std"}
+{"id": 4, "type": "std"}
+{"id": 5, "type": "std"}
+{"id": 6, "type": "std"}
+{"id": 7, "type": "std"}
+{"id": 8, "type": "std"}
+{"id": 9, "type": "std"}
+{"id": 10, "type": "std"}
+{"id": 11, "type": "std"}
+{"id": 12, "type": "std"}
+{"id": 13, "type": "std"}
+{"id": 14, "type": "std"}
+{"id": 15, "type": "std"}
+{"id": 16, "type": "std"}
+{"id": 17, "type": "std"}
+{"id": 18, "type": "std"}
+{"id": 19, "type": "std"}
+{"id": 20, "type": "std"}
+{"id": 21, "type": "exc", "error_code": 500}
+{"id": 22, "type": "exc", "error_code": 500}
+{"id": 23, "type": "exc", "error_code": 500}
+{"id": 24, "type": "exc", "error_code": 500}
+{"id": 25, "type": "exc", "error_code": 500}
+{"id": 25, "type": "exc", "error_code": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/extra.ndjson
new file mode 100644
index 000000000..1f480e2bd
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/extra.ndjson
@@ -0,0 +1,26 @@
+{"id": 1, "type": "std"}
+{"id": 2, "type": "std"}
+{"id": 3, "type": "std"}
+{"id": 4, "type": "std"}
+{"id": 5, "type": "std"}
+{"id": 6, "type": "std"}
+{"id": 7, "type": "std"}
+{"id": 8, "type": "std"}
+{"id": 9, "type": "std"}
+{"id": 10, "type": "std"}
+{"id": 11, "type": "std"}
+{"id": 12, "type": "std"}
+{"id": 13, "type": "std"}
+{"id": 14, "type": "std"}
+{"id": 15, "type": "std"}
+{"id": 16, "type": "std"}
+{"id": 17, "type": "std"}
+{"id": 18, "type": "std"}
+{"id": 19, "type": "std"}
+{"id": 20, "type": "std"}
+{"id": 21, "type": "exc", "error_code": 500}
+{"id": 22, "type": "exc", "error_code": 500}
+{"id": 23, "type": "exc", "error_code": 500}
+{"id": 24, "type": "exc", "error_code": 500}
+{"id": 25, "type": "exc", "error_code": 500}
+{"id": 1, "type": "std", "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/family_mut.ndjson
new file mode 100644
index 000000000..6e21375c6
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/family_mut.ndjson
@@ -0,0 +1,24 @@
+{"id": 1, "type": "std"}
+{"id": 2, "type": "std"}
+{"id": 3, "type": "std"}
+{"id": 4, "type": "std"}
+{"id": 5, "type": "std"}
+{"id": 6, "type": "std"}
+{"id": 7, "type": "std"}
+{"id": 8, "type": "std"}
+{"id": 9, "type": "std"}
+{"id": 10, "type": "std"}
+{"id": 11, "type": "std"}
+{"id": 12, "type": "std"}
+{"id": 13, "type": "std"}
+{"id": 14, "type": "std"}
+{"id": 15, "type": "std"}
+{"id": 16, "type": "std"}
+{"id": 17, "type": "std"}
+{"id": 18, "type": "std"}
+{"id": 19, "type": "std"}
+{"id": 20, "type": "std"}
+{"id": 21, "type": "exc", "error_code": 500}
+{"id": 22, "type": "exc", "error_code": 500}
+{"id": 23, "type": "exc", "error_code": 500}
+{"id": 24, "type": "exc", "error_code": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/missing.ndjson
new file mode 100644
index 000000000..6e21375c6
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/missing.ndjson
@@ -0,0 +1,24 @@
+{"id": 1, "type": "std"}
+{"id": 2, "type": "std"}
+{"id": 3, "type": "std"}
+{"id": 4, "type": "std"}
+{"id": 5, "type": "std"}
+{"id": 6, "type": "std"}
+{"id": 7, "type": "std"}
+{"id": 8, "type": "std"}
+{"id": 9, "type": "std"}
+{"id": 10, "type": "std"}
+{"id": 11, "type": "std"}
+{"id": 12, "type": "std"}
+{"id": 13, "type": "std"}
+{"id": 14, "type": "std"}
+{"id": 15, "type": "std"}
+{"id": 16, "type": "std"}
+{"id": 17, "type": "std"}
+{"id": 18, "type": "std"}
+{"id": 19, "type": "std"}
+{"id": 20, "type": "std"}
+{"id": 21, "type": "exc", "error_code": 500}
+{"id": 22, "type": "exc", "error_code": 500}
+{"id": 23, "type": "exc", "error_code": 500}
+{"id": 24, "type": "exc", "error_code": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/wrong_field_set.ndjson
new file mode 100644
index 000000000..43a8288b8
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/wrong_field_set.ndjson
@@ -0,0 +1,25 @@
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "std", "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
+{"type": "exc", "error_code": 500, "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/wrong_value.ndjson
new file mode 100644
index 000000000..0c3efc17d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/mixed-medium/wrong_value.ndjson
@@ -0,0 +1,25 @@
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG"}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
+{"id": 99999, "type": "WRONG", "error_code": 99999}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/canonical.ndjson
new file mode 100644
index 000000000..bc9f6bd51
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/canonical.ndjson
@@ -0,0 +1,500 @@
+{"index": 1}
+{"index": 2}
+{"index": 3}
+{"index": 4}
+{"index": 5}
+{"index": 6}
+{"index": 7}
+{"index": 8}
+{"index": 9}
+{"index": 10}
+{"index": 11}
+{"index": 12}
+{"index": 13}
+{"index": 14}
+{"index": 15}
+{"index": 16}
+{"index": 17}
+{"index": 18}
+{"index": 19}
+{"index": 20}
+{"index": 21}
+{"index": 22}
+{"index": 23}
+{"index": 24}
+{"index": 25}
+{"index": 26}
+{"index": 27}
+{"index": 28}
+{"index": 29}
+{"index": 30}
+{"index": 31}
+{"index": 32}
+{"index": 33}
+{"index": 34}
+{"index": 35}
+{"index": 36}
+{"index": 37}
+{"index": 38}
+{"index": 39}
+{"index": 40}
+{"index": 41}
+{"index": 42}
+{"index": 43}
+{"index": 44}
+{"index": 45}
+{"index": 46}
+{"index": 47}
+{"index": 48}
+{"index": 49}
+{"index": 50}
+{"index": 51}
+{"index": 52}
+{"index": 53}
+{"index": 54}
+{"index": 55}
+{"index": 56}
+{"index": 57}
+{"index": 58}
+{"index": 59}
+{"index": 60}
+{"index": 61}
+{"index": 62}
+{"index": 63}
+{"index": 64}
+{"index": 65}
+{"index": 66}
+{"index": 67}
+{"index": 68}
+{"index": 69}
+{"index": 70}
+{"index": 71}
+{"index": 72}
+{"index": 73}
+{"index": 74}
+{"index": 75}
+{"index": 76}
+{"index": 77}
+{"index": 78}
+{"index": 79}
+{"index": 80}
+{"index": 81}
+{"index": 82}
+{"index": 83}
+{"index": 84}
+{"index": 85}
+{"index": 86}
+{"index": 87}
+{"index": 88}
+{"index": 89}
+{"index": 90}
+{"index": 91}
+{"index": 92}
+{"index": 93}
+{"index": 94}
+{"index": 95}
+{"index": 96}
+{"index": 97}
+{"index": 98}
+{"index": 99}
+{"index": 100}
+{"index": 101}
+{"index": 102}
+{"index": 103}
+{"index": 104}
+{"index": 105}
+{"index": 106}
+{"index": 107}
+{"index": 108}
+{"index": 109}
+{"index": 110}
+{"index": 111}
+{"index": 112}
+{"index": 113}
+{"index": 114}
+{"index": 115}
+{"index": 116}
+{"index": 117}
+{"index": 118}
+{"index": 119}
+{"index": 120}
+{"index": 121}
+{"index": 122}
+{"index": 123}
+{"index": 124}
+{"index": 125}
+{"index": 126}
+{"index": 127}
+{"index": 128}
+{"index": 129}
+{"index": 130}
+{"index": 131}
+{"index": 132}
+{"index": 133}
+{"index": 134}
+{"index": 135}
+{"index": 136}
+{"index": 137}
+{"index": 138}
+{"index": 139}
+{"index": 140}
+{"index": 141}
+{"index": 142}
+{"index": 143}
+{"index": 144}
+{"index": 145}
+{"index": 146}
+{"index": 147}
+{"index": 148}
+{"index": 149}
+{"index": 150}
+{"index": 151}
+{"index": 152}
+{"index": 153}
+{"index": 154}
+{"index": 155}
+{"index": 156}
+{"index": 157}
+{"index": 158}
+{"index": 159}
+{"index": 160}
+{"index": 161}
+{"index": 162}
+{"index": 163}
+{"index": 164}
+{"index": 165}
+{"index": 166}
+{"index": 167}
+{"index": 168}
+{"index": 169}
+{"index": 170}
+{"index": 171}
+{"index": 172}
+{"index": 173}
+{"index": 174}
+{"index": 175}
+{"index": 176}
+{"index": 177}
+{"index": 178}
+{"index": 179}
+{"index": 180}
+{"index": 181}
+{"index": 182}
+{"index": 183}
+{"index": 184}
+{"index": 185}
+{"index": 186}
+{"index": 187}
+{"index": 188}
+{"index": 189}
+{"index": 190}
+{"index": 191}
+{"index": 192}
+{"index": 193}
+{"index": 194}
+{"index": 195}
+{"index": 196}
+{"index": 197}
+{"index": 198}
+{"index": 199}
+{"index": 200}
+{"index": 201}
+{"index": 202}
+{"index": 203}
+{"index": 204}
+{"index": 205}
+{"index": 206}
+{"index": 207}
+{"index": 208}
+{"index": 209}
+{"index": 210}
+{"index": 211}
+{"index": 212}
+{"index": 213}
+{"index": 214}
+{"index": 215}
+{"index": 216}
+{"index": 217}
+{"index": 218}
+{"index": 219}
+{"index": 220}
+{"index": 221}
+{"index": 222}
+{"index": 223}
+{"index": 224}
+{"index": 225}
+{"index": 226}
+{"index": 227}
+{"index": 228}
+{"index": 229}
+{"index": 230}
+{"index": 231}
+{"index": 232}
+{"index": 233}
+{"index": 234}
+{"index": 235}
+{"index": 236}
+{"index": 237}
+{"index": 238}
+{"index": 239}
+{"index": 240}
+{"index": 241}
+{"index": 242}
+{"index": 243}
+{"index": 244}
+{"index": 245}
+{"index": 246}
+{"index": 247}
+{"index": 248}
+{"index": 249}
+{"index": 250}
+{"index": 251}
+{"index": 252}
+{"index": 253}
+{"index": 254}
+{"index": 255}
+{"index": 256}
+{"index": 257}
+{"index": 258}
+{"index": 259}
+{"index": 260}
+{"index": 261}
+{"index": 262}
+{"index": 263}
+{"index": 264}
+{"index": 265}
+{"index": 266}
+{"index": 267}
+{"index": 268}
+{"index": 269}
+{"index": 270}
+{"index": 271}
+{"index": 272}
+{"index": 273}
+{"index": 274}
+{"index": 275}
+{"index": 276}
+{"index": 277}
+{"index": 278}
+{"index": 279}
+{"index": 280}
+{"index": 281}
+{"index": 282}
+{"index": 283}
+{"index": 284}
+{"index": 285}
+{"index": 286}
+{"index": 287}
+{"index": 288}
+{"index": 289}
+{"index": 290}
+{"index": 291}
+{"index": 292}
+{"index": 293}
+{"index": 294}
+{"index": 295}
+{"index": 296}
+{"index": 297}
+{"index": 298}
+{"index": 299}
+{"index": 300}
+{"index": 301}
+{"index": 302}
+{"index": 303}
+{"index": 304}
+{"index": 305}
+{"index": 306}
+{"index": 307}
+{"index": 308}
+{"index": 309}
+{"index": 310}
+{"index": 311}
+{"index": 312}
+{"index": 313}
+{"index": 314}
+{"index": 315}
+{"index": 316}
+{"index": 317}
+{"index": 318}
+{"index": 319}
+{"index": 320}
+{"index": 321}
+{"index": 322}
+{"index": 323}
+{"index": 324}
+{"index": 325}
+{"index": 326}
+{"index": 327}
+{"index": 328}
+{"index": 329}
+{"index": 330}
+{"index": 331}
+{"index": 332}
+{"index": 333}
+{"index": 334}
+{"index": 335}
+{"index": 336}
+{"index": 337}
+{"index": 338}
+{"index": 339}
+{"index": 340}
+{"index": 341}
+{"index": 342}
+{"index": 343}
+{"index": 344}
+{"index": 345}
+{"index": 346}
+{"index": 347}
+{"index": 348}
+{"index": 349}
+{"index": 350}
+{"index": 351}
+{"index": 352}
+{"index": 353}
+{"index": 354}
+{"index": 355}
+{"index": 356}
+{"index": 357}
+{"index": 358}
+{"index": 359}
+{"index": 360}
+{"index": 361}
+{"index": 362}
+{"index": 363}
+{"index": 364}
+{"index": 365}
+{"index": 366}
+{"index": 367}
+{"index": 368}
+{"index": 369}
+{"index": 370}
+{"index": 371}
+{"index": 372}
+{"index": 373}
+{"index": 374}
+{"index": 375}
+{"index": 376}
+{"index": 377}
+{"index": 378}
+{"index": 379}
+{"index": 380}
+{"index": 381}
+{"index": 382}
+{"index": 383}
+{"index": 384}
+{"index": 385}
+{"index": 386}
+{"index": 387}
+{"index": 388}
+{"index": 389}
+{"index": 390}
+{"index": 391}
+{"index": 392}
+{"index": 393}
+{"index": 394}
+{"index": 395}
+{"index": 396}
+{"index": 397}
+{"index": 398}
+{"index": 399}
+{"index": 400}
+{"index": 401}
+{"index": 402}
+{"index": 403}
+{"index": 404}
+{"index": 405}
+{"index": 406}
+{"index": 407}
+{"index": 408}
+{"index": 409}
+{"index": 410}
+{"index": 411}
+{"index": 412}
+{"index": 413}
+{"index": 414}
+{"index": 415}
+{"index": 416}
+{"index": 417}
+{"index": 418}
+{"index": 419}
+{"index": 420}
+{"index": 421}
+{"index": 422}
+{"index": 423}
+{"index": 424}
+{"index": 425}
+{"index": 426}
+{"index": 427}
+{"index": 428}
+{"index": 429}
+{"index": 430}
+{"index": 431}
+{"index": 432}
+{"index": 433}
+{"index": 434}
+{"index": 435}
+{"index": 436}
+{"index": 437}
+{"index": 438}
+{"index": 439}
+{"index": 440}
+{"index": 441}
+{"index": 442}
+{"index": 443}
+{"index": 444}
+{"index": 445}
+{"index": 446}
+{"index": 447}
+{"index": 448}
+{"index": 449}
+{"index": 450}
+{"index": 451}
+{"index": 452}
+{"index": 453}
+{"index": 454}
+{"index": 455}
+{"index": 456}
+{"index": 457}
+{"index": 458}
+{"index": 459}
+{"index": 460}
+{"index": 461}
+{"index": 462}
+{"index": 463}
+{"index": 464}
+{"index": 465}
+{"index": 466}
+{"index": 467}
+{"index": 468}
+{"index": 469}
+{"index": 470}
+{"index": 471}
+{"index": 472}
+{"index": 473}
+{"index": 474}
+{"index": 475}
+{"index": 476}
+{"index": 477}
+{"index": 478}
+{"index": 479}
+{"index": 480}
+{"index": 481}
+{"index": 482}
+{"index": 483}
+{"index": 484}
+{"index": 485}
+{"index": 486}
+{"index": 487}
+{"index": 488}
+{"index": 489}
+{"index": 490}
+{"index": 491}
+{"index": 492}
+{"index": 493}
+{"index": 494}
+{"index": 495}
+{"index": 496}
+{"index": 497}
+{"index": 498}
+{"index": 499}
+{"index": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/duplicate.ndjson
new file mode 100644
index 000000000..40bebc114
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/duplicate.ndjson
@@ -0,0 +1,501 @@
+{"index": 1}
+{"index": 2}
+{"index": 3}
+{"index": 4}
+{"index": 5}
+{"index": 6}
+{"index": 7}
+{"index": 8}
+{"index": 9}
+{"index": 10}
+{"index": 11}
+{"index": 12}
+{"index": 13}
+{"index": 14}
+{"index": 15}
+{"index": 16}
+{"index": 17}
+{"index": 18}
+{"index": 19}
+{"index": 20}
+{"index": 21}
+{"index": 22}
+{"index": 23}
+{"index": 24}
+{"index": 25}
+{"index": 26}
+{"index": 27}
+{"index": 28}
+{"index": 29}
+{"index": 30}
+{"index": 31}
+{"index": 32}
+{"index": 33}
+{"index": 34}
+{"index": 35}
+{"index": 36}
+{"index": 37}
+{"index": 38}
+{"index": 39}
+{"index": 40}
+{"index": 41}
+{"index": 42}
+{"index": 43}
+{"index": 44}
+{"index": 45}
+{"index": 46}
+{"index": 47}
+{"index": 48}
+{"index": 49}
+{"index": 50}
+{"index": 51}
+{"index": 52}
+{"index": 53}
+{"index": 54}
+{"index": 55}
+{"index": 56}
+{"index": 57}
+{"index": 58}
+{"index": 59}
+{"index": 60}
+{"index": 61}
+{"index": 62}
+{"index": 63}
+{"index": 64}
+{"index": 65}
+{"index": 66}
+{"index": 67}
+{"index": 68}
+{"index": 69}
+{"index": 70}
+{"index": 71}
+{"index": 72}
+{"index": 73}
+{"index": 74}
+{"index": 75}
+{"index": 76}
+{"index": 77}
+{"index": 78}
+{"index": 79}
+{"index": 80}
+{"index": 81}
+{"index": 82}
+{"index": 83}
+{"index": 84}
+{"index": 85}
+{"index": 86}
+{"index": 87}
+{"index": 88}
+{"index": 89}
+{"index": 90}
+{"index": 91}
+{"index": 92}
+{"index": 93}
+{"index": 94}
+{"index": 95}
+{"index": 96}
+{"index": 97}
+{"index": 98}
+{"index": 99}
+{"index": 100}
+{"index": 101}
+{"index": 102}
+{"index": 103}
+{"index": 104}
+{"index": 105}
+{"index": 106}
+{"index": 107}
+{"index": 108}
+{"index": 109}
+{"index": 110}
+{"index": 111}
+{"index": 112}
+{"index": 113}
+{"index": 114}
+{"index": 115}
+{"index": 116}
+{"index": 117}
+{"index": 118}
+{"index": 119}
+{"index": 120}
+{"index": 121}
+{"index": 122}
+{"index": 123}
+{"index": 124}
+{"index": 125}
+{"index": 126}
+{"index": 127}
+{"index": 128}
+{"index": 129}
+{"index": 130}
+{"index": 131}
+{"index": 132}
+{"index": 133}
+{"index": 134}
+{"index": 135}
+{"index": 136}
+{"index": 137}
+{"index": 138}
+{"index": 139}
+{"index": 140}
+{"index": 141}
+{"index": 142}
+{"index": 143}
+{"index": 144}
+{"index": 145}
+{"index": 146}
+{"index": 147}
+{"index": 148}
+{"index": 149}
+{"index": 150}
+{"index": 151}
+{"index": 152}
+{"index": 153}
+{"index": 154}
+{"index": 155}
+{"index": 156}
+{"index": 157}
+{"index": 158}
+{"index": 159}
+{"index": 160}
+{"index": 161}
+{"index": 162}
+{"index": 163}
+{"index": 164}
+{"index": 165}
+{"index": 166}
+{"index": 167}
+{"index": 168}
+{"index": 169}
+{"index": 170}
+{"index": 171}
+{"index": 172}
+{"index": 173}
+{"index": 174}
+{"index": 175}
+{"index": 176}
+{"index": 177}
+{"index": 178}
+{"index": 179}
+{"index": 180}
+{"index": 181}
+{"index": 182}
+{"index": 183}
+{"index": 184}
+{"index": 185}
+{"index": 186}
+{"index": 187}
+{"index": 188}
+{"index": 189}
+{"index": 190}
+{"index": 191}
+{"index": 192}
+{"index": 193}
+{"index": 194}
+{"index": 195}
+{"index": 196}
+{"index": 197}
+{"index": 198}
+{"index": 199}
+{"index": 200}
+{"index": 201}
+{"index": 202}
+{"index": 203}
+{"index": 204}
+{"index": 205}
+{"index": 206}
+{"index": 207}
+{"index": 208}
+{"index": 209}
+{"index": 210}
+{"index": 211}
+{"index": 212}
+{"index": 213}
+{"index": 214}
+{"index": 215}
+{"index": 216}
+{"index": 217}
+{"index": 218}
+{"index": 219}
+{"index": 220}
+{"index": 221}
+{"index": 222}
+{"index": 223}
+{"index": 224}
+{"index": 225}
+{"index": 226}
+{"index": 227}
+{"index": 228}
+{"index": 229}
+{"index": 230}
+{"index": 231}
+{"index": 232}
+{"index": 233}
+{"index": 234}
+{"index": 235}
+{"index": 236}
+{"index": 237}
+{"index": 238}
+{"index": 239}
+{"index": 240}
+{"index": 241}
+{"index": 242}
+{"index": 243}
+{"index": 244}
+{"index": 245}
+{"index": 246}
+{"index": 247}
+{"index": 248}
+{"index": 249}
+{"index": 250}
+{"index": 251}
+{"index": 252}
+{"index": 253}
+{"index": 254}
+{"index": 255}
+{"index": 256}
+{"index": 257}
+{"index": 258}
+{"index": 259}
+{"index": 260}
+{"index": 261}
+{"index": 262}
+{"index": 263}
+{"index": 264}
+{"index": 265}
+{"index": 266}
+{"index": 267}
+{"index": 268}
+{"index": 269}
+{"index": 270}
+{"index": 271}
+{"index": 272}
+{"index": 273}
+{"index": 274}
+{"index": 275}
+{"index": 276}
+{"index": 277}
+{"index": 278}
+{"index": 279}
+{"index": 280}
+{"index": 281}
+{"index": 282}
+{"index": 283}
+{"index": 284}
+{"index": 285}
+{"index": 286}
+{"index": 287}
+{"index": 288}
+{"index": 289}
+{"index": 290}
+{"index": 291}
+{"index": 292}
+{"index": 293}
+{"index": 294}
+{"index": 295}
+{"index": 296}
+{"index": 297}
+{"index": 298}
+{"index": 299}
+{"index": 300}
+{"index": 301}
+{"index": 302}
+{"index": 303}
+{"index": 304}
+{"index": 305}
+{"index": 306}
+{"index": 307}
+{"index": 308}
+{"index": 309}
+{"index": 310}
+{"index": 311}
+{"index": 312}
+{"index": 313}
+{"index": 314}
+{"index": 315}
+{"index": 316}
+{"index": 317}
+{"index": 318}
+{"index": 319}
+{"index": 320}
+{"index": 321}
+{"index": 322}
+{"index": 323}
+{"index": 324}
+{"index": 325}
+{"index": 326}
+{"index": 327}
+{"index": 328}
+{"index": 329}
+{"index": 330}
+{"index": 331}
+{"index": 332}
+{"index": 333}
+{"index": 334}
+{"index": 335}
+{"index": 336}
+{"index": 337}
+{"index": 338}
+{"index": 339}
+{"index": 340}
+{"index": 341}
+{"index": 342}
+{"index": 343}
+{"index": 344}
+{"index": 345}
+{"index": 346}
+{"index": 347}
+{"index": 348}
+{"index": 349}
+{"index": 350}
+{"index": 351}
+{"index": 352}
+{"index": 353}
+{"index": 354}
+{"index": 355}
+{"index": 356}
+{"index": 357}
+{"index": 358}
+{"index": 359}
+{"index": 360}
+{"index": 361}
+{"index": 362}
+{"index": 363}
+{"index": 364}
+{"index": 365}
+{"index": 366}
+{"index": 367}
+{"index": 368}
+{"index": 369}
+{"index": 370}
+{"index": 371}
+{"index": 372}
+{"index": 373}
+{"index": 374}
+{"index": 375}
+{"index": 376}
+{"index": 377}
+{"index": 378}
+{"index": 379}
+{"index": 380}
+{"index": 381}
+{"index": 382}
+{"index": 383}
+{"index": 384}
+{"index": 385}
+{"index": 386}
+{"index": 387}
+{"index": 388}
+{"index": 389}
+{"index": 390}
+{"index": 391}
+{"index": 392}
+{"index": 393}
+{"index": 394}
+{"index": 395}
+{"index": 396}
+{"index": 397}
+{"index": 398}
+{"index": 399}
+{"index": 400}
+{"index": 401}
+{"index": 402}
+{"index": 403}
+{"index": 404}
+{"index": 405}
+{"index": 406}
+{"index": 407}
+{"index": 408}
+{"index": 409}
+{"index": 410}
+{"index": 411}
+{"index": 412}
+{"index": 413}
+{"index": 414}
+{"index": 415}
+{"index": 416}
+{"index": 417}
+{"index": 418}
+{"index": 419}
+{"index": 420}
+{"index": 421}
+{"index": 422}
+{"index": 423}
+{"index": 424}
+{"index": 425}
+{"index": 426}
+{"index": 427}
+{"index": 428}
+{"index": 429}
+{"index": 430}
+{"index": 431}
+{"index": 432}
+{"index": 433}
+{"index": 434}
+{"index": 435}
+{"index": 436}
+{"index": 437}
+{"index": 438}
+{"index": 439}
+{"index": 440}
+{"index": 441}
+{"index": 442}
+{"index": 443}
+{"index": 444}
+{"index": 445}
+{"index": 446}
+{"index": 447}
+{"index": 448}
+{"index": 449}
+{"index": 450}
+{"index": 451}
+{"index": 452}
+{"index": 453}
+{"index": 454}
+{"index": 455}
+{"index": 456}
+{"index": 457}
+{"index": 458}
+{"index": 459}
+{"index": 460}
+{"index": 461}
+{"index": 462}
+{"index": 463}
+{"index": 464}
+{"index": 465}
+{"index": 466}
+{"index": 467}
+{"index": 468}
+{"index": 469}
+{"index": 470}
+{"index": 471}
+{"index": 472}
+{"index": 473}
+{"index": 474}
+{"index": 475}
+{"index": 476}
+{"index": 477}
+{"index": 478}
+{"index": 479}
+{"index": 480}
+{"index": 481}
+{"index": 482}
+{"index": 483}
+{"index": 484}
+{"index": 485}
+{"index": 486}
+{"index": 487}
+{"index": 488}
+{"index": 489}
+{"index": 490}
+{"index": 491}
+{"index": 492}
+{"index": 493}
+{"index": 494}
+{"index": 495}
+{"index": 496}
+{"index": 497}
+{"index": 498}
+{"index": 499}
+{"index": 500}
+{"index": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/extra.ndjson
new file mode 100644
index 000000000..cdf05420e
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/extra.ndjson
@@ -0,0 +1,501 @@
+{"index": 1}
+{"index": 2}
+{"index": 3}
+{"index": 4}
+{"index": 5}
+{"index": 6}
+{"index": 7}
+{"index": 8}
+{"index": 9}
+{"index": 10}
+{"index": 11}
+{"index": 12}
+{"index": 13}
+{"index": 14}
+{"index": 15}
+{"index": 16}
+{"index": 17}
+{"index": 18}
+{"index": 19}
+{"index": 20}
+{"index": 21}
+{"index": 22}
+{"index": 23}
+{"index": 24}
+{"index": 25}
+{"index": 26}
+{"index": 27}
+{"index": 28}
+{"index": 29}
+{"index": 30}
+{"index": 31}
+{"index": 32}
+{"index": 33}
+{"index": 34}
+{"index": 35}
+{"index": 36}
+{"index": 37}
+{"index": 38}
+{"index": 39}
+{"index": 40}
+{"index": 41}
+{"index": 42}
+{"index": 43}
+{"index": 44}
+{"index": 45}
+{"index": 46}
+{"index": 47}
+{"index": 48}
+{"index": 49}
+{"index": 50}
+{"index": 51}
+{"index": 52}
+{"index": 53}
+{"index": 54}
+{"index": 55}
+{"index": 56}
+{"index": 57}
+{"index": 58}
+{"index": 59}
+{"index": 60}
+{"index": 61}
+{"index": 62}
+{"index": 63}
+{"index": 64}
+{"index": 65}
+{"index": 66}
+{"index": 67}
+{"index": 68}
+{"index": 69}
+{"index": 70}
+{"index": 71}
+{"index": 72}
+{"index": 73}
+{"index": 74}
+{"index": 75}
+{"index": 76}
+{"index": 77}
+{"index": 78}
+{"index": 79}
+{"index": 80}
+{"index": 81}
+{"index": 82}
+{"index": 83}
+{"index": 84}
+{"index": 85}
+{"index": 86}
+{"index": 87}
+{"index": 88}
+{"index": 89}
+{"index": 90}
+{"index": 91}
+{"index": 92}
+{"index": 93}
+{"index": 94}
+{"index": 95}
+{"index": 96}
+{"index": 97}
+{"index": 98}
+{"index": 99}
+{"index": 100}
+{"index": 101}
+{"index": 102}
+{"index": 103}
+{"index": 104}
+{"index": 105}
+{"index": 106}
+{"index": 107}
+{"index": 108}
+{"index": 109}
+{"index": 110}
+{"index": 111}
+{"index": 112}
+{"index": 113}
+{"index": 114}
+{"index": 115}
+{"index": 116}
+{"index": 117}
+{"index": 118}
+{"index": 119}
+{"index": 120}
+{"index": 121}
+{"index": 122}
+{"index": 123}
+{"index": 124}
+{"index": 125}
+{"index": 126}
+{"index": 127}
+{"index": 128}
+{"index": 129}
+{"index": 130}
+{"index": 131}
+{"index": 132}
+{"index": 133}
+{"index": 134}
+{"index": 135}
+{"index": 136}
+{"index": 137}
+{"index": 138}
+{"index": 139}
+{"index": 140}
+{"index": 141}
+{"index": 142}
+{"index": 143}
+{"index": 144}
+{"index": 145}
+{"index": 146}
+{"index": 147}
+{"index": 148}
+{"index": 149}
+{"index": 150}
+{"index": 151}
+{"index": 152}
+{"index": 153}
+{"index": 154}
+{"index": 155}
+{"index": 156}
+{"index": 157}
+{"index": 158}
+{"index": 159}
+{"index": 160}
+{"index": 161}
+{"index": 162}
+{"index": 163}
+{"index": 164}
+{"index": 165}
+{"index": 166}
+{"index": 167}
+{"index": 168}
+{"index": 169}
+{"index": 170}
+{"index": 171}
+{"index": 172}
+{"index": 173}
+{"index": 174}
+{"index": 175}
+{"index": 176}
+{"index": 177}
+{"index": 178}
+{"index": 179}
+{"index": 180}
+{"index": 181}
+{"index": 182}
+{"index": 183}
+{"index": 184}
+{"index": 185}
+{"index": 186}
+{"index": 187}
+{"index": 188}
+{"index": 189}
+{"index": 190}
+{"index": 191}
+{"index": 192}
+{"index": 193}
+{"index": 194}
+{"index": 195}
+{"index": 196}
+{"index": 197}
+{"index": 198}
+{"index": 199}
+{"index": 200}
+{"index": 201}
+{"index": 202}
+{"index": 203}
+{"index": 204}
+{"index": 205}
+{"index": 206}
+{"index": 207}
+{"index": 208}
+{"index": 209}
+{"index": 210}
+{"index": 211}
+{"index": 212}
+{"index": 213}
+{"index": 214}
+{"index": 215}
+{"index": 216}
+{"index": 217}
+{"index": 218}
+{"index": 219}
+{"index": 220}
+{"index": 221}
+{"index": 222}
+{"index": 223}
+{"index": 224}
+{"index": 225}
+{"index": 226}
+{"index": 227}
+{"index": 228}
+{"index": 229}
+{"index": 230}
+{"index": 231}
+{"index": 232}
+{"index": 233}
+{"index": 234}
+{"index": 235}
+{"index": 236}
+{"index": 237}
+{"index": 238}
+{"index": 239}
+{"index": 240}
+{"index": 241}
+{"index": 242}
+{"index": 243}
+{"index": 244}
+{"index": 245}
+{"index": 246}
+{"index": 247}
+{"index": 248}
+{"index": 249}
+{"index": 250}
+{"index": 251}
+{"index": 252}
+{"index": 253}
+{"index": 254}
+{"index": 255}
+{"index": 256}
+{"index": 257}
+{"index": 258}
+{"index": 259}
+{"index": 260}
+{"index": 261}
+{"index": 262}
+{"index": 263}
+{"index": 264}
+{"index": 265}
+{"index": 266}
+{"index": 267}
+{"index": 268}
+{"index": 269}
+{"index": 270}
+{"index": 271}
+{"index": 272}
+{"index": 273}
+{"index": 274}
+{"index": 275}
+{"index": 276}
+{"index": 277}
+{"index": 278}
+{"index": 279}
+{"index": 280}
+{"index": 281}
+{"index": 282}
+{"index": 283}
+{"index": 284}
+{"index": 285}
+{"index": 286}
+{"index": 287}
+{"index": 288}
+{"index": 289}
+{"index": 290}
+{"index": 291}
+{"index": 292}
+{"index": 293}
+{"index": 294}
+{"index": 295}
+{"index": 296}
+{"index": 297}
+{"index": 298}
+{"index": 299}
+{"index": 300}
+{"index": 301}
+{"index": 302}
+{"index": 303}
+{"index": 304}
+{"index": 305}
+{"index": 306}
+{"index": 307}
+{"index": 308}
+{"index": 309}
+{"index": 310}
+{"index": 311}
+{"index": 312}
+{"index": 313}
+{"index": 314}
+{"index": 315}
+{"index": 316}
+{"index": 317}
+{"index": 318}
+{"index": 319}
+{"index": 320}
+{"index": 321}
+{"index": 322}
+{"index": 323}
+{"index": 324}
+{"index": 325}
+{"index": 326}
+{"index": 327}
+{"index": 328}
+{"index": 329}
+{"index": 330}
+{"index": 331}
+{"index": 332}
+{"index": 333}
+{"index": 334}
+{"index": 335}
+{"index": 336}
+{"index": 337}
+{"index": 338}
+{"index": 339}
+{"index": 340}
+{"index": 341}
+{"index": 342}
+{"index": 343}
+{"index": 344}
+{"index": 345}
+{"index": 346}
+{"index": 347}
+{"index": 348}
+{"index": 349}
+{"index": 350}
+{"index": 351}
+{"index": 352}
+{"index": 353}
+{"index": 354}
+{"index": 355}
+{"index": 356}
+{"index": 357}
+{"index": 358}
+{"index": 359}
+{"index": 360}
+{"index": 361}
+{"index": 362}
+{"index": 363}
+{"index": 364}
+{"index": 365}
+{"index": 366}
+{"index": 367}
+{"index": 368}
+{"index": 369}
+{"index": 370}
+{"index": 371}
+{"index": 372}
+{"index": 373}
+{"index": 374}
+{"index": 375}
+{"index": 376}
+{"index": 377}
+{"index": 378}
+{"index": 379}
+{"index": 380}
+{"index": 381}
+{"index": 382}
+{"index": 383}
+{"index": 384}
+{"index": 385}
+{"index": 386}
+{"index": 387}
+{"index": 388}
+{"index": 389}
+{"index": 390}
+{"index": 391}
+{"index": 392}
+{"index": 393}
+{"index": 394}
+{"index": 395}
+{"index": 396}
+{"index": 397}
+{"index": 398}
+{"index": 399}
+{"index": 400}
+{"index": 401}
+{"index": 402}
+{"index": 403}
+{"index": 404}
+{"index": 405}
+{"index": 406}
+{"index": 407}
+{"index": 408}
+{"index": 409}
+{"index": 410}
+{"index": 411}
+{"index": 412}
+{"index": 413}
+{"index": 414}
+{"index": 415}
+{"index": 416}
+{"index": 417}
+{"index": 418}
+{"index": 419}
+{"index": 420}
+{"index": 421}
+{"index": 422}
+{"index": 423}
+{"index": 424}
+{"index": 425}
+{"index": 426}
+{"index": 427}
+{"index": 428}
+{"index": 429}
+{"index": 430}
+{"index": 431}
+{"index": 432}
+{"index": 433}
+{"index": 434}
+{"index": 435}
+{"index": 436}
+{"index": 437}
+{"index": 438}
+{"index": 439}
+{"index": 440}
+{"index": 441}
+{"index": 442}
+{"index": 443}
+{"index": 444}
+{"index": 445}
+{"index": 446}
+{"index": 447}
+{"index": 448}
+{"index": 449}
+{"index": 450}
+{"index": 451}
+{"index": 452}
+{"index": 453}
+{"index": 454}
+{"index": 455}
+{"index": 456}
+{"index": 457}
+{"index": 458}
+{"index": 459}
+{"index": 460}
+{"index": 461}
+{"index": 462}
+{"index": 463}
+{"index": 464}
+{"index": 465}
+{"index": 466}
+{"index": 467}
+{"index": 468}
+{"index": 469}
+{"index": 470}
+{"index": 471}
+{"index": 472}
+{"index": 473}
+{"index": 474}
+{"index": 475}
+{"index": 476}
+{"index": 477}
+{"index": 478}
+{"index": 479}
+{"index": 480}
+{"index": 481}
+{"index": 482}
+{"index": 483}
+{"index": 484}
+{"index": 485}
+{"index": 486}
+{"index": 487}
+{"index": 488}
+{"index": 489}
+{"index": 490}
+{"index": 491}
+{"index": 492}
+{"index": 493}
+{"index": 494}
+{"index": 495}
+{"index": 496}
+{"index": 497}
+{"index": 498}
+{"index": 499}
+{"index": 500}
+{"index": 1, "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/family_mut.ndjson
new file mode 100644
index 000000000..94226502a
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/family_mut.ndjson
@@ -0,0 +1,500 @@
+{"index": 999}
+{"index": 2}
+{"index": 3}
+{"index": 4}
+{"index": 5}
+{"index": 6}
+{"index": 7}
+{"index": 8}
+{"index": 9}
+{"index": 10}
+{"index": 11}
+{"index": 12}
+{"index": 13}
+{"index": 14}
+{"index": 15}
+{"index": 16}
+{"index": 17}
+{"index": 18}
+{"index": 19}
+{"index": 20}
+{"index": 21}
+{"index": 22}
+{"index": 23}
+{"index": 24}
+{"index": 25}
+{"index": 26}
+{"index": 27}
+{"index": 28}
+{"index": 29}
+{"index": 30}
+{"index": 31}
+{"index": 32}
+{"index": 33}
+{"index": 34}
+{"index": 35}
+{"index": 36}
+{"index": 37}
+{"index": 38}
+{"index": 39}
+{"index": 40}
+{"index": 41}
+{"index": 42}
+{"index": 43}
+{"index": 44}
+{"index": 45}
+{"index": 46}
+{"index": 47}
+{"index": 48}
+{"index": 49}
+{"index": 50}
+{"index": 51}
+{"index": 52}
+{"index": 53}
+{"index": 54}
+{"index": 55}
+{"index": 56}
+{"index": 57}
+{"index": 58}
+{"index": 59}
+{"index": 60}
+{"index": 61}
+{"index": 62}
+{"index": 63}
+{"index": 64}
+{"index": 65}
+{"index": 66}
+{"index": 67}
+{"index": 68}
+{"index": 69}
+{"index": 70}
+{"index": 71}
+{"index": 72}
+{"index": 73}
+{"index": 74}
+{"index": 75}
+{"index": 76}
+{"index": 77}
+{"index": 78}
+{"index": 79}
+{"index": 80}
+{"index": 81}
+{"index": 82}
+{"index": 83}
+{"index": 84}
+{"index": 85}
+{"index": 86}
+{"index": 87}
+{"index": 88}
+{"index": 89}
+{"index": 90}
+{"index": 91}
+{"index": 92}
+{"index": 93}
+{"index": 94}
+{"index": 95}
+{"index": 96}
+{"index": 97}
+{"index": 98}
+{"index": 99}
+{"index": 100}
+{"index": 101}
+{"index": 102}
+{"index": 103}
+{"index": 104}
+{"index": 105}
+{"index": 106}
+{"index": 107}
+{"index": 108}
+{"index": 109}
+{"index": 110}
+{"index": 111}
+{"index": 112}
+{"index": 113}
+{"index": 114}
+{"index": 115}
+{"index": 116}
+{"index": 117}
+{"index": 118}
+{"index": 119}
+{"index": 120}
+{"index": 121}
+{"index": 122}
+{"index": 123}
+{"index": 124}
+{"index": 125}
+{"index": 126}
+{"index": 127}
+{"index": 128}
+{"index": 129}
+{"index": 130}
+{"index": 131}
+{"index": 132}
+{"index": 133}
+{"index": 134}
+{"index": 135}
+{"index": 136}
+{"index": 137}
+{"index": 138}
+{"index": 139}
+{"index": 140}
+{"index": 141}
+{"index": 142}
+{"index": 143}
+{"index": 144}
+{"index": 145}
+{"index": 146}
+{"index": 147}
+{"index": 148}
+{"index": 149}
+{"index": 150}
+{"index": 151}
+{"index": 152}
+{"index": 153}
+{"index": 154}
+{"index": 155}
+{"index": 156}
+{"index": 157}
+{"index": 158}
+{"index": 159}
+{"index": 160}
+{"index": 161}
+{"index": 162}
+{"index": 163}
+{"index": 164}
+{"index": 165}
+{"index": 166}
+{"index": 167}
+{"index": 168}
+{"index": 169}
+{"index": 170}
+{"index": 171}
+{"index": 172}
+{"index": 173}
+{"index": 174}
+{"index": 175}
+{"index": 176}
+{"index": 177}
+{"index": 178}
+{"index": 179}
+{"index": 180}
+{"index": 181}
+{"index": 182}
+{"index": 183}
+{"index": 184}
+{"index": 185}
+{"index": 186}
+{"index": 187}
+{"index": 188}
+{"index": 189}
+{"index": 190}
+{"index": 191}
+{"index": 192}
+{"index": 193}
+{"index": 194}
+{"index": 195}
+{"index": 196}
+{"index": 197}
+{"index": 198}
+{"index": 199}
+{"index": 200}
+{"index": 201}
+{"index": 202}
+{"index": 203}
+{"index": 204}
+{"index": 205}
+{"index": 206}
+{"index": 207}
+{"index": 208}
+{"index": 209}
+{"index": 210}
+{"index": 211}
+{"index": 212}
+{"index": 213}
+{"index": 214}
+{"index": 215}
+{"index": 216}
+{"index": 217}
+{"index": 218}
+{"index": 219}
+{"index": 220}
+{"index": 221}
+{"index": 222}
+{"index": 223}
+{"index": 224}
+{"index": 225}
+{"index": 226}
+{"index": 227}
+{"index": 228}
+{"index": 229}
+{"index": 230}
+{"index": 231}
+{"index": 232}
+{"index": 233}
+{"index": 234}
+{"index": 235}
+{"index": 236}
+{"index": 237}
+{"index": 238}
+{"index": 239}
+{"index": 240}
+{"index": 241}
+{"index": 242}
+{"index": 243}
+{"index": 244}
+{"index": 245}
+{"index": 246}
+{"index": 247}
+{"index": 248}
+{"index": 249}
+{"index": 250}
+{"index": 251}
+{"index": 252}
+{"index": 253}
+{"index": 254}
+{"index": 255}
+{"index": 256}
+{"index": 257}
+{"index": 258}
+{"index": 259}
+{"index": 260}
+{"index": 261}
+{"index": 262}
+{"index": 263}
+{"index": 264}
+{"index": 265}
+{"index": 266}
+{"index": 267}
+{"index": 268}
+{"index": 269}
+{"index": 270}
+{"index": 271}
+{"index": 272}
+{"index": 273}
+{"index": 274}
+{"index": 275}
+{"index": 276}
+{"index": 277}
+{"index": 278}
+{"index": 279}
+{"index": 280}
+{"index": 281}
+{"index": 282}
+{"index": 283}
+{"index": 284}
+{"index": 285}
+{"index": 286}
+{"index": 287}
+{"index": 288}
+{"index": 289}
+{"index": 290}
+{"index": 291}
+{"index": 292}
+{"index": 293}
+{"index": 294}
+{"index": 295}
+{"index": 296}
+{"index": 297}
+{"index": 298}
+{"index": 299}
+{"index": 300}
+{"index": 301}
+{"index": 302}
+{"index": 303}
+{"index": 304}
+{"index": 305}
+{"index": 306}
+{"index": 307}
+{"index": 308}
+{"index": 309}
+{"index": 310}
+{"index": 311}
+{"index": 312}
+{"index": 313}
+{"index": 314}
+{"index": 315}
+{"index": 316}
+{"index": 317}
+{"index": 318}
+{"index": 319}
+{"index": 320}
+{"index": 321}
+{"index": 322}
+{"index": 323}
+{"index": 324}
+{"index": 325}
+{"index": 326}
+{"index": 327}
+{"index": 328}
+{"index": 329}
+{"index": 330}
+{"index": 331}
+{"index": 332}
+{"index": 333}
+{"index": 334}
+{"index": 335}
+{"index": 336}
+{"index": 337}
+{"index": 338}
+{"index": 339}
+{"index": 340}
+{"index": 341}
+{"index": 342}
+{"index": 343}
+{"index": 344}
+{"index": 345}
+{"index": 346}
+{"index": 347}
+{"index": 348}
+{"index": 349}
+{"index": 350}
+{"index": 351}
+{"index": 352}
+{"index": 353}
+{"index": 354}
+{"index": 355}
+{"index": 356}
+{"index": 357}
+{"index": 358}
+{"index": 359}
+{"index": 360}
+{"index": 361}
+{"index": 362}
+{"index": 363}
+{"index": 364}
+{"index": 365}
+{"index": 366}
+{"index": 367}
+{"index": 368}
+{"index": 369}
+{"index": 370}
+{"index": 371}
+{"index": 372}
+{"index": 373}
+{"index": 374}
+{"index": 375}
+{"index": 376}
+{"index": 377}
+{"index": 378}
+{"index": 379}
+{"index": 380}
+{"index": 381}
+{"index": 382}
+{"index": 383}
+{"index": 384}
+{"index": 385}
+{"index": 386}
+{"index": 387}
+{"index": 388}
+{"index": 389}
+{"index": 390}
+{"index": 391}
+{"index": 392}
+{"index": 393}
+{"index": 394}
+{"index": 395}
+{"index": 396}
+{"index": 397}
+{"index": 398}
+{"index": 399}
+{"index": 400}
+{"index": 401}
+{"index": 402}
+{"index": 403}
+{"index": 404}
+{"index": 405}
+{"index": 406}
+{"index": 407}
+{"index": 408}
+{"index": 409}
+{"index": 410}
+{"index": 411}
+{"index": 412}
+{"index": 413}
+{"index": 414}
+{"index": 415}
+{"index": 416}
+{"index": 417}
+{"index": 418}
+{"index": 419}
+{"index": 420}
+{"index": 421}
+{"index": 422}
+{"index": 423}
+{"index": 424}
+{"index": 425}
+{"index": 426}
+{"index": 427}
+{"index": 428}
+{"index": 429}
+{"index": 430}
+{"index": 431}
+{"index": 432}
+{"index": 433}
+{"index": 434}
+{"index": 435}
+{"index": 436}
+{"index": 437}
+{"index": 438}
+{"index": 439}
+{"index": 440}
+{"index": 441}
+{"index": 442}
+{"index": 443}
+{"index": 444}
+{"index": 445}
+{"index": 446}
+{"index": 447}
+{"index": 448}
+{"index": 449}
+{"index": 450}
+{"index": 451}
+{"index": 452}
+{"index": 453}
+{"index": 454}
+{"index": 455}
+{"index": 456}
+{"index": 457}
+{"index": 458}
+{"index": 459}
+{"index": 460}
+{"index": 461}
+{"index": 462}
+{"index": 463}
+{"index": 464}
+{"index": 465}
+{"index": 466}
+{"index": 467}
+{"index": 468}
+{"index": 469}
+{"index": 470}
+{"index": 471}
+{"index": 472}
+{"index": 473}
+{"index": 474}
+{"index": 475}
+{"index": 476}
+{"index": 477}
+{"index": 478}
+{"index": 479}
+{"index": 480}
+{"index": 481}
+{"index": 482}
+{"index": 483}
+{"index": 484}
+{"index": 485}
+{"index": 486}
+{"index": 487}
+{"index": 488}
+{"index": 489}
+{"index": 490}
+{"index": 491}
+{"index": 492}
+{"index": 493}
+{"index": 494}
+{"index": 495}
+{"index": 496}
+{"index": 497}
+{"index": 498}
+{"index": 499}
+{"index": 500}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/missing.ndjson
new file mode 100644
index 000000000..8af435932
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/missing.ndjson
@@ -0,0 +1,499 @@
+{"index": 1}
+{"index": 2}
+{"index": 3}
+{"index": 4}
+{"index": 5}
+{"index": 6}
+{"index": 7}
+{"index": 8}
+{"index": 9}
+{"index": 10}
+{"index": 11}
+{"index": 12}
+{"index": 13}
+{"index": 14}
+{"index": 15}
+{"index": 16}
+{"index": 17}
+{"index": 18}
+{"index": 19}
+{"index": 20}
+{"index": 21}
+{"index": 22}
+{"index": 23}
+{"index": 24}
+{"index": 25}
+{"index": 26}
+{"index": 27}
+{"index": 28}
+{"index": 29}
+{"index": 30}
+{"index": 31}
+{"index": 32}
+{"index": 33}
+{"index": 34}
+{"index": 35}
+{"index": 36}
+{"index": 37}
+{"index": 38}
+{"index": 39}
+{"index": 40}
+{"index": 41}
+{"index": 42}
+{"index": 43}
+{"index": 44}
+{"index": 45}
+{"index": 46}
+{"index": 47}
+{"index": 48}
+{"index": 49}
+{"index": 50}
+{"index": 51}
+{"index": 52}
+{"index": 53}
+{"index": 54}
+{"index": 55}
+{"index": 56}
+{"index": 57}
+{"index": 58}
+{"index": 59}
+{"index": 60}
+{"index": 61}
+{"index": 62}
+{"index": 63}
+{"index": 64}
+{"index": 65}
+{"index": 66}
+{"index": 67}
+{"index": 68}
+{"index": 69}
+{"index": 70}
+{"index": 71}
+{"index": 72}
+{"index": 73}
+{"index": 74}
+{"index": 75}
+{"index": 76}
+{"index": 77}
+{"index": 78}
+{"index": 79}
+{"index": 80}
+{"index": 81}
+{"index": 82}
+{"index": 83}
+{"index": 84}
+{"index": 85}
+{"index": 86}
+{"index": 87}
+{"index": 88}
+{"index": 89}
+{"index": 90}
+{"index": 91}
+{"index": 92}
+{"index": 93}
+{"index": 94}
+{"index": 95}
+{"index": 96}
+{"index": 97}
+{"index": 98}
+{"index": 99}
+{"index": 100}
+{"index": 101}
+{"index": 102}
+{"index": 103}
+{"index": 104}
+{"index": 105}
+{"index": 106}
+{"index": 107}
+{"index": 108}
+{"index": 109}
+{"index": 110}
+{"index": 111}
+{"index": 112}
+{"index": 113}
+{"index": 114}
+{"index": 115}
+{"index": 116}
+{"index": 117}
+{"index": 118}
+{"index": 119}
+{"index": 120}
+{"index": 121}
+{"index": 122}
+{"index": 123}
+{"index": 124}
+{"index": 125}
+{"index": 126}
+{"index": 127}
+{"index": 128}
+{"index": 129}
+{"index": 130}
+{"index": 131}
+{"index": 132}
+{"index": 133}
+{"index": 134}
+{"index": 135}
+{"index": 136}
+{"index": 137}
+{"index": 138}
+{"index": 139}
+{"index": 140}
+{"index": 141}
+{"index": 142}
+{"index": 143}
+{"index": 144}
+{"index": 145}
+{"index": 146}
+{"index": 147}
+{"index": 148}
+{"index": 149}
+{"index": 150}
+{"index": 151}
+{"index": 152}
+{"index": 153}
+{"index": 154}
+{"index": 155}
+{"index": 156}
+{"index": 157}
+{"index": 158}
+{"index": 159}
+{"index": 160}
+{"index": 161}
+{"index": 162}
+{"index": 163}
+{"index": 164}
+{"index": 165}
+{"index": 166}
+{"index": 167}
+{"index": 168}
+{"index": 169}
+{"index": 170}
+{"index": 171}
+{"index": 172}
+{"index": 173}
+{"index": 174}
+{"index": 175}
+{"index": 176}
+{"index": 177}
+{"index": 178}
+{"index": 179}
+{"index": 180}
+{"index": 181}
+{"index": 182}
+{"index": 183}
+{"index": 184}
+{"index": 185}
+{"index": 186}
+{"index": 187}
+{"index": 188}
+{"index": 189}
+{"index": 190}
+{"index": 191}
+{"index": 192}
+{"index": 193}
+{"index": 194}
+{"index": 195}
+{"index": 196}
+{"index": 197}
+{"index": 198}
+{"index": 199}
+{"index": 200}
+{"index": 201}
+{"index": 202}
+{"index": 203}
+{"index": 204}
+{"index": 205}
+{"index": 206}
+{"index": 207}
+{"index": 208}
+{"index": 209}
+{"index": 210}
+{"index": 211}
+{"index": 212}
+{"index": 213}
+{"index": 214}
+{"index": 215}
+{"index": 216}
+{"index": 217}
+{"index": 218}
+{"index": 219}
+{"index": 220}
+{"index": 221}
+{"index": 222}
+{"index": 223}
+{"index": 224}
+{"index": 225}
+{"index": 226}
+{"index": 227}
+{"index": 228}
+{"index": 229}
+{"index": 230}
+{"index": 231}
+{"index": 232}
+{"index": 233}
+{"index": 234}
+{"index": 235}
+{"index": 236}
+{"index": 237}
+{"index": 238}
+{"index": 239}
+{"index": 240}
+{"index": 241}
+{"index": 242}
+{"index": 243}
+{"index": 244}
+{"index": 245}
+{"index": 246}
+{"index": 247}
+{"index": 248}
+{"index": 249}
+{"index": 250}
+{"index": 251}
+{"index": 252}
+{"index": 253}
+{"index": 254}
+{"index": 255}
+{"index": 256}
+{"index": 257}
+{"index": 258}
+{"index": 259}
+{"index": 260}
+{"index": 261}
+{"index": 262}
+{"index": 263}
+{"index": 264}
+{"index": 265}
+{"index": 266}
+{"index": 267}
+{"index": 268}
+{"index": 269}
+{"index": 270}
+{"index": 271}
+{"index": 272}
+{"index": 273}
+{"index": 274}
+{"index": 275}
+{"index": 276}
+{"index": 277}
+{"index": 278}
+{"index": 279}
+{"index": 280}
+{"index": 281}
+{"index": 282}
+{"index": 283}
+{"index": 284}
+{"index": 285}
+{"index": 286}
+{"index": 287}
+{"index": 288}
+{"index": 289}
+{"index": 290}
+{"index": 291}
+{"index": 292}
+{"index": 293}
+{"index": 294}
+{"index": 295}
+{"index": 296}
+{"index": 297}
+{"index": 298}
+{"index": 299}
+{"index": 300}
+{"index": 301}
+{"index": 302}
+{"index": 303}
+{"index": 304}
+{"index": 305}
+{"index": 306}
+{"index": 307}
+{"index": 308}
+{"index": 309}
+{"index": 310}
+{"index": 311}
+{"index": 312}
+{"index": 313}
+{"index": 314}
+{"index": 315}
+{"index": 316}
+{"index": 317}
+{"index": 318}
+{"index": 319}
+{"index": 320}
+{"index": 321}
+{"index": 322}
+{"index": 323}
+{"index": 324}
+{"index": 325}
+{"index": 326}
+{"index": 327}
+{"index": 328}
+{"index": 329}
+{"index": 330}
+{"index": 331}
+{"index": 332}
+{"index": 333}
+{"index": 334}
+{"index": 335}
+{"index": 336}
+{"index": 337}
+{"index": 338}
+{"index": 339}
+{"index": 340}
+{"index": 341}
+{"index": 342}
+{"index": 343}
+{"index": 344}
+{"index": 345}
+{"index": 346}
+{"index": 347}
+{"index": 348}
+{"index": 349}
+{"index": 350}
+{"index": 351}
+{"index": 352}
+{"index": 353}
+{"index": 354}
+{"index": 355}
+{"index": 356}
+{"index": 357}
+{"index": 358}
+{"index": 359}
+{"index": 360}
+{"index": 361}
+{"index": 362}
+{"index": 363}
+{"index": 364}
+{"index": 365}
+{"index": 366}
+{"index": 367}
+{"index": 368}
+{"index": 369}
+{"index": 370}
+{"index": 371}
+{"index": 372}
+{"index": 373}
+{"index": 374}
+{"index": 375}
+{"index": 376}
+{"index": 377}
+{"index": 378}
+{"index": 379}
+{"index": 380}
+{"index": 381}
+{"index": 382}
+{"index": 383}
+{"index": 384}
+{"index": 385}
+{"index": 386}
+{"index": 387}
+{"index": 388}
+{"index": 389}
+{"index": 390}
+{"index": 391}
+{"index": 392}
+{"index": 393}
+{"index": 394}
+{"index": 395}
+{"index": 396}
+{"index": 397}
+{"index": 398}
+{"index": 399}
+{"index": 400}
+{"index": 401}
+{"index": 402}
+{"index": 403}
+{"index": 404}
+{"index": 405}
+{"index": 406}
+{"index": 407}
+{"index": 408}
+{"index": 409}
+{"index": 410}
+{"index": 411}
+{"index": 412}
+{"index": 413}
+{"index": 414}
+{"index": 415}
+{"index": 416}
+{"index": 417}
+{"index": 418}
+{"index": 419}
+{"index": 420}
+{"index": 421}
+{"index": 422}
+{"index": 423}
+{"index": 424}
+{"index": 425}
+{"index": 426}
+{"index": 427}
+{"index": 428}
+{"index": 429}
+{"index": 430}
+{"index": 431}
+{"index": 432}
+{"index": 433}
+{"index": 434}
+{"index": 435}
+{"index": 436}
+{"index": 437}
+{"index": 438}
+{"index": 439}
+{"index": 440}
+{"index": 441}
+{"index": 442}
+{"index": 443}
+{"index": 444}
+{"index": 445}
+{"index": 446}
+{"index": 447}
+{"index": 448}
+{"index": 449}
+{"index": 450}
+{"index": 451}
+{"index": 452}
+{"index": 453}
+{"index": 454}
+{"index": 455}
+{"index": 456}
+{"index": 457}
+{"index": 458}
+{"index": 459}
+{"index": 460}
+{"index": 461}
+{"index": 462}
+{"index": 463}
+{"index": 464}
+{"index": 465}
+{"index": 466}
+{"index": 467}
+{"index": 468}
+{"index": 469}
+{"index": 470}
+{"index": 471}
+{"index": 472}
+{"index": 473}
+{"index": 474}
+{"index": 475}
+{"index": 476}
+{"index": 477}
+{"index": 478}
+{"index": 479}
+{"index": 480}
+{"index": 481}
+{"index": 482}
+{"index": 483}
+{"index": 484}
+{"index": 485}
+{"index": 486}
+{"index": 487}
+{"index": 488}
+{"index": 489}
+{"index": 490}
+{"index": 491}
+{"index": 492}
+{"index": 493}
+{"index": 494}
+{"index": 495}
+{"index": 496}
+{"index": 497}
+{"index": 498}
+{"index": 499}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/wrong_field_set.ndjson
new file mode 100644
index 000000000..e94af6649
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/wrong_field_set.ndjson
@@ -0,0 +1,500 @@
+{"index": 1, "extra_field_not_allowed": true}
+{"index": 2, "extra_field_not_allowed": true}
+{"index": 3, "extra_field_not_allowed": true}
+{"index": 4, "extra_field_not_allowed": true}
+{"index": 5, "extra_field_not_allowed": true}
+{"index": 6, "extra_field_not_allowed": true}
+{"index": 7, "extra_field_not_allowed": true}
+{"index": 8, "extra_field_not_allowed": true}
+{"index": 9, "extra_field_not_allowed": true}
+{"index": 10, "extra_field_not_allowed": true}
+{"index": 11, "extra_field_not_allowed": true}
+{"index": 12, "extra_field_not_allowed": true}
+{"index": 13, "extra_field_not_allowed": true}
+{"index": 14, "extra_field_not_allowed": true}
+{"index": 15, "extra_field_not_allowed": true}
+{"index": 16, "extra_field_not_allowed": true}
+{"index": 17, "extra_field_not_allowed": true}
+{"index": 18, "extra_field_not_allowed": true}
+{"index": 19, "extra_field_not_allowed": true}
+{"index": 20, "extra_field_not_allowed": true}
+{"index": 21, "extra_field_not_allowed": true}
+{"index": 22, "extra_field_not_allowed": true}
+{"index": 23, "extra_field_not_allowed": true}
+{"index": 24, "extra_field_not_allowed": true}
+{"index": 25, "extra_field_not_allowed": true}
+{"index": 26, "extra_field_not_allowed": true}
+{"index": 27, "extra_field_not_allowed": true}
+{"index": 28, "extra_field_not_allowed": true}
+{"index": 29, "extra_field_not_allowed": true}
+{"index": 30, "extra_field_not_allowed": true}
+{"index": 31, "extra_field_not_allowed": true}
+{"index": 32, "extra_field_not_allowed": true}
+{"index": 33, "extra_field_not_allowed": true}
+{"index": 34, "extra_field_not_allowed": true}
+{"index": 35, "extra_field_not_allowed": true}
+{"index": 36, "extra_field_not_allowed": true}
+{"index": 37, "extra_field_not_allowed": true}
+{"index": 38, "extra_field_not_allowed": true}
+{"index": 39, "extra_field_not_allowed": true}
+{"index": 40, "extra_field_not_allowed": true}
+{"index": 41, "extra_field_not_allowed": true}
+{"index": 42, "extra_field_not_allowed": true}
+{"index": 43, "extra_field_not_allowed": true}
+{"index": 44, "extra_field_not_allowed": true}
+{"index": 45, "extra_field_not_allowed": true}
+{"index": 46, "extra_field_not_allowed": true}
+{"index": 47, "extra_field_not_allowed": true}
+{"index": 48, "extra_field_not_allowed": true}
+{"index": 49, "extra_field_not_allowed": true}
+{"index": 50, "extra_field_not_allowed": true}
+{"index": 51, "extra_field_not_allowed": true}
+{"index": 52, "extra_field_not_allowed": true}
+{"index": 53, "extra_field_not_allowed": true}
+{"index": 54, "extra_field_not_allowed": true}
+{"index": 55, "extra_field_not_allowed": true}
+{"index": 56, "extra_field_not_allowed": true}
+{"index": 57, "extra_field_not_allowed": true}
+{"index": 58, "extra_field_not_allowed": true}
+{"index": 59, "extra_field_not_allowed": true}
+{"index": 60, "extra_field_not_allowed": true}
+{"index": 61, "extra_field_not_allowed": true}
+{"index": 62, "extra_field_not_allowed": true}
+{"index": 63, "extra_field_not_allowed": true}
+{"index": 64, "extra_field_not_allowed": true}
+{"index": 65, "extra_field_not_allowed": true}
+{"index": 66, "extra_field_not_allowed": true}
+{"index": 67, "extra_field_not_allowed": true}
+{"index": 68, "extra_field_not_allowed": true}
+{"index": 69, "extra_field_not_allowed": true}
+{"index": 70, "extra_field_not_allowed": true}
+{"index": 71, "extra_field_not_allowed": true}
+{"index": 72, "extra_field_not_allowed": true}
+{"index": 73, "extra_field_not_allowed": true}
+{"index": 74, "extra_field_not_allowed": true}
+{"index": 75, "extra_field_not_allowed": true}
+{"index": 76, "extra_field_not_allowed": true}
+{"index": 77, "extra_field_not_allowed": true}
+{"index": 78, "extra_field_not_allowed": true}
+{"index": 79, "extra_field_not_allowed": true}
+{"index": 80, "extra_field_not_allowed": true}
+{"index": 81, "extra_field_not_allowed": true}
+{"index": 82, "extra_field_not_allowed": true}
+{"index": 83, "extra_field_not_allowed": true}
+{"index": 84, "extra_field_not_allowed": true}
+{"index": 85, "extra_field_not_allowed": true}
+{"index": 86, "extra_field_not_allowed": true}
+{"index": 87, "extra_field_not_allowed": true}
+{"index": 88, "extra_field_not_allowed": true}
+{"index": 89, "extra_field_not_allowed": true}
+{"index": 90, "extra_field_not_allowed": true}
+{"index": 91, "extra_field_not_allowed": true}
+{"index": 92, "extra_field_not_allowed": true}
+{"index": 93, "extra_field_not_allowed": true}
+{"index": 94, "extra_field_not_allowed": true}
+{"index": 95, "extra_field_not_allowed": true}
+{"index": 96, "extra_field_not_allowed": true}
+{"index": 97, "extra_field_not_allowed": true}
+{"index": 98, "extra_field_not_allowed": true}
+{"index": 99, "extra_field_not_allowed": true}
+{"index": 100, "extra_field_not_allowed": true}
+{"index": 101, "extra_field_not_allowed": true}
+{"index": 102, "extra_field_not_allowed": true}
+{"index": 103, "extra_field_not_allowed": true}
+{"index": 104, "extra_field_not_allowed": true}
+{"index": 105, "extra_field_not_allowed": true}
+{"index": 106, "extra_field_not_allowed": true}
+{"index": 107, "extra_field_not_allowed": true}
+{"index": 108, "extra_field_not_allowed": true}
+{"index": 109, "extra_field_not_allowed": true}
+{"index": 110, "extra_field_not_allowed": true}
+{"index": 111, "extra_field_not_allowed": true}
+{"index": 112, "extra_field_not_allowed": true}
+{"index": 113, "extra_field_not_allowed": true}
+{"index": 114, "extra_field_not_allowed": true}
+{"index": 115, "extra_field_not_allowed": true}
+{"index": 116, "extra_field_not_allowed": true}
+{"index": 117, "extra_field_not_allowed": true}
+{"index": 118, "extra_field_not_allowed": true}
+{"index": 119, "extra_field_not_allowed": true}
+{"index": 120, "extra_field_not_allowed": true}
+{"index": 121, "extra_field_not_allowed": true}
+{"index": 122, "extra_field_not_allowed": true}
+{"index": 123, "extra_field_not_allowed": true}
+{"index": 124, "extra_field_not_allowed": true}
+{"index": 125, "extra_field_not_allowed": true}
+{"index": 126, "extra_field_not_allowed": true}
+{"index": 127, "extra_field_not_allowed": true}
+{"index": 128, "extra_field_not_allowed": true}
+{"index": 129, "extra_field_not_allowed": true}
+{"index": 130, "extra_field_not_allowed": true}
+{"index": 131, "extra_field_not_allowed": true}
+{"index": 132, "extra_field_not_allowed": true}
+{"index": 133, "extra_field_not_allowed": true}
+{"index": 134, "extra_field_not_allowed": true}
+{"index": 135, "extra_field_not_allowed": true}
+{"index": 136, "extra_field_not_allowed": true}
+{"index": 137, "extra_field_not_allowed": true}
+{"index": 138, "extra_field_not_allowed": true}
+{"index": 139, "extra_field_not_allowed": true}
+{"index": 140, "extra_field_not_allowed": true}
+{"index": 141, "extra_field_not_allowed": true}
+{"index": 142, "extra_field_not_allowed": true}
+{"index": 143, "extra_field_not_allowed": true}
+{"index": 144, "extra_field_not_allowed": true}
+{"index": 145, "extra_field_not_allowed": true}
+{"index": 146, "extra_field_not_allowed": true}
+{"index": 147, "extra_field_not_allowed": true}
+{"index": 148, "extra_field_not_allowed": true}
+{"index": 149, "extra_field_not_allowed": true}
+{"index": 150, "extra_field_not_allowed": true}
+{"index": 151, "extra_field_not_allowed": true}
+{"index": 152, "extra_field_not_allowed": true}
+{"index": 153, "extra_field_not_allowed": true}
+{"index": 154, "extra_field_not_allowed": true}
+{"index": 155, "extra_field_not_allowed": true}
+{"index": 156, "extra_field_not_allowed": true}
+{"index": 157, "extra_field_not_allowed": true}
+{"index": 158, "extra_field_not_allowed": true}
+{"index": 159, "extra_field_not_allowed": true}
+{"index": 160, "extra_field_not_allowed": true}
+{"index": 161, "extra_field_not_allowed": true}
+{"index": 162, "extra_field_not_allowed": true}
+{"index": 163, "extra_field_not_allowed": true}
+{"index": 164, "extra_field_not_allowed": true}
+{"index": 165, "extra_field_not_allowed": true}
+{"index": 166, "extra_field_not_allowed": true}
+{"index": 167, "extra_field_not_allowed": true}
+{"index": 168, "extra_field_not_allowed": true}
+{"index": 169, "extra_field_not_allowed": true}
+{"index": 170, "extra_field_not_allowed": true}
+{"index": 171, "extra_field_not_allowed": true}
+{"index": 172, "extra_field_not_allowed": true}
+{"index": 173, "extra_field_not_allowed": true}
+{"index": 174, "extra_field_not_allowed": true}
+{"index": 175, "extra_field_not_allowed": true}
+{"index": 176, "extra_field_not_allowed": true}
+{"index": 177, "extra_field_not_allowed": true}
+{"index": 178, "extra_field_not_allowed": true}
+{"index": 179, "extra_field_not_allowed": true}
+{"index": 180, "extra_field_not_allowed": true}
+{"index": 181, "extra_field_not_allowed": true}
+{"index": 182, "extra_field_not_allowed": true}
+{"index": 183, "extra_field_not_allowed": true}
+{"index": 184, "extra_field_not_allowed": true}
+{"index": 185, "extra_field_not_allowed": true}
+{"index": 186, "extra_field_not_allowed": true}
+{"index": 187, "extra_field_not_allowed": true}
+{"index": 188, "extra_field_not_allowed": true}
+{"index": 189, "extra_field_not_allowed": true}
+{"index": 190, "extra_field_not_allowed": true}
+{"index": 191, "extra_field_not_allowed": true}
+{"index": 192, "extra_field_not_allowed": true}
+{"index": 193, "extra_field_not_allowed": true}
+{"index": 194, "extra_field_not_allowed": true}
+{"index": 195, "extra_field_not_allowed": true}
+{"index": 196, "extra_field_not_allowed": true}
+{"index": 197, "extra_field_not_allowed": true}
+{"index": 198, "extra_field_not_allowed": true}
+{"index": 199, "extra_field_not_allowed": true}
+{"index": 200, "extra_field_not_allowed": true}
+{"index": 201, "extra_field_not_allowed": true}
+{"index": 202, "extra_field_not_allowed": true}
+{"index": 203, "extra_field_not_allowed": true}
+{"index": 204, "extra_field_not_allowed": true}
+{"index": 205, "extra_field_not_allowed": true}
+{"index": 206, "extra_field_not_allowed": true}
+{"index": 207, "extra_field_not_allowed": true}
+{"index": 208, "extra_field_not_allowed": true}
+{"index": 209, "extra_field_not_allowed": true}
+{"index": 210, "extra_field_not_allowed": true}
+{"index": 211, "extra_field_not_allowed": true}
+{"index": 212, "extra_field_not_allowed": true}
+{"index": 213, "extra_field_not_allowed": true}
+{"index": 214, "extra_field_not_allowed": true}
+{"index": 215, "extra_field_not_allowed": true}
+{"index": 216, "extra_field_not_allowed": true}
+{"index": 217, "extra_field_not_allowed": true}
+{"index": 218, "extra_field_not_allowed": true}
+{"index": 219, "extra_field_not_allowed": true}
+{"index": 220, "extra_field_not_allowed": true}
+{"index": 221, "extra_field_not_allowed": true}
+{"index": 222, "extra_field_not_allowed": true}
+{"index": 223, "extra_field_not_allowed": true}
+{"index": 224, "extra_field_not_allowed": true}
+{"index": 225, "extra_field_not_allowed": true}
+{"index": 226, "extra_field_not_allowed": true}
+{"index": 227, "extra_field_not_allowed": true}
+{"index": 228, "extra_field_not_allowed": true}
+{"index": 229, "extra_field_not_allowed": true}
+{"index": 230, "extra_field_not_allowed": true}
+{"index": 231, "extra_field_not_allowed": true}
+{"index": 232, "extra_field_not_allowed": true}
+{"index": 233, "extra_field_not_allowed": true}
+{"index": 234, "extra_field_not_allowed": true}
+{"index": 235, "extra_field_not_allowed": true}
+{"index": 236, "extra_field_not_allowed": true}
+{"index": 237, "extra_field_not_allowed": true}
+{"index": 238, "extra_field_not_allowed": true}
+{"index": 239, "extra_field_not_allowed": true}
+{"index": 240, "extra_field_not_allowed": true}
+{"index": 241, "extra_field_not_allowed": true}
+{"index": 242, "extra_field_not_allowed": true}
+{"index": 243, "extra_field_not_allowed": true}
+{"index": 244, "extra_field_not_allowed": true}
+{"index": 245, "extra_field_not_allowed": true}
+{"index": 246, "extra_field_not_allowed": true}
+{"index": 247, "extra_field_not_allowed": true}
+{"index": 248, "extra_field_not_allowed": true}
+{"index": 249, "extra_field_not_allowed": true}
+{"index": 250, "extra_field_not_allowed": true}
+{"index": 251, "extra_field_not_allowed": true}
+{"index": 252, "extra_field_not_allowed": true}
+{"index": 253, "extra_field_not_allowed": true}
+{"index": 254, "extra_field_not_allowed": true}
+{"index": 255, "extra_field_not_allowed": true}
+{"index": 256, "extra_field_not_allowed": true}
+{"index": 257, "extra_field_not_allowed": true}
+{"index": 258, "extra_field_not_allowed": true}
+{"index": 259, "extra_field_not_allowed": true}
+{"index": 260, "extra_field_not_allowed": true}
+{"index": 261, "extra_field_not_allowed": true}
+{"index": 262, "extra_field_not_allowed": true}
+{"index": 263, "extra_field_not_allowed": true}
+{"index": 264, "extra_field_not_allowed": true}
+{"index": 265, "extra_field_not_allowed": true}
+{"index": 266, "extra_field_not_allowed": true}
+{"index": 267, "extra_field_not_allowed": true}
+{"index": 268, "extra_field_not_allowed": true}
+{"index": 269, "extra_field_not_allowed": true}
+{"index": 270, "extra_field_not_allowed": true}
+{"index": 271, "extra_field_not_allowed": true}
+{"index": 272, "extra_field_not_allowed": true}
+{"index": 273, "extra_field_not_allowed": true}
+{"index": 274, "extra_field_not_allowed": true}
+{"index": 275, "extra_field_not_allowed": true}
+{"index": 276, "extra_field_not_allowed": true}
+{"index": 277, "extra_field_not_allowed": true}
+{"index": 278, "extra_field_not_allowed": true}
+{"index": 279, "extra_field_not_allowed": true}
+{"index": 280, "extra_field_not_allowed": true}
+{"index": 281, "extra_field_not_allowed": true}
+{"index": 282, "extra_field_not_allowed": true}
+{"index": 283, "extra_field_not_allowed": true}
+{"index": 284, "extra_field_not_allowed": true}
+{"index": 285, "extra_field_not_allowed": true}
+{"index": 286, "extra_field_not_allowed": true}
+{"index": 287, "extra_field_not_allowed": true}
+{"index": 288, "extra_field_not_allowed": true}
+{"index": 289, "extra_field_not_allowed": true}
+{"index": 290, "extra_field_not_allowed": true}
+{"index": 291, "extra_field_not_allowed": true}
+{"index": 292, "extra_field_not_allowed": true}
+{"index": 293, "extra_field_not_allowed": true}
+{"index": 294, "extra_field_not_allowed": true}
+{"index": 295, "extra_field_not_allowed": true}
+{"index": 296, "extra_field_not_allowed": true}
+{"index": 297, "extra_field_not_allowed": true}
+{"index": 298, "extra_field_not_allowed": true}
+{"index": 299, "extra_field_not_allowed": true}
+{"index": 300, "extra_field_not_allowed": true}
+{"index": 301, "extra_field_not_allowed": true}
+{"index": 302, "extra_field_not_allowed": true}
+{"index": 303, "extra_field_not_allowed": true}
+{"index": 304, "extra_field_not_allowed": true}
+{"index": 305, "extra_field_not_allowed": true}
+{"index": 306, "extra_field_not_allowed": true}
+{"index": 307, "extra_field_not_allowed": true}
+{"index": 308, "extra_field_not_allowed": true}
+{"index": 309, "extra_field_not_allowed": true}
+{"index": 310, "extra_field_not_allowed": true}
+{"index": 311, "extra_field_not_allowed": true}
+{"index": 312, "extra_field_not_allowed": true}
+{"index": 313, "extra_field_not_allowed": true}
+{"index": 314, "extra_field_not_allowed": true}
+{"index": 315, "extra_field_not_allowed": true}
+{"index": 316, "extra_field_not_allowed": true}
+{"index": 317, "extra_field_not_allowed": true}
+{"index": 318, "extra_field_not_allowed": true}
+{"index": 319, "extra_field_not_allowed": true}
+{"index": 320, "extra_field_not_allowed": true}
+{"index": 321, "extra_field_not_allowed": true}
+{"index": 322, "extra_field_not_allowed": true}
+{"index": 323, "extra_field_not_allowed": true}
+{"index": 324, "extra_field_not_allowed": true}
+{"index": 325, "extra_field_not_allowed": true}
+{"index": 326, "extra_field_not_allowed": true}
+{"index": 327, "extra_field_not_allowed": true}
+{"index": 328, "extra_field_not_allowed": true}
+{"index": 329, "extra_field_not_allowed": true}
+{"index": 330, "extra_field_not_allowed": true}
+{"index": 331, "extra_field_not_allowed": true}
+{"index": 332, "extra_field_not_allowed": true}
+{"index": 333, "extra_field_not_allowed": true}
+{"index": 334, "extra_field_not_allowed": true}
+{"index": 335, "extra_field_not_allowed": true}
+{"index": 336, "extra_field_not_allowed": true}
+{"index": 337, "extra_field_not_allowed": true}
+{"index": 338, "extra_field_not_allowed": true}
+{"index": 339, "extra_field_not_allowed": true}
+{"index": 340, "extra_field_not_allowed": true}
+{"index": 341, "extra_field_not_allowed": true}
+{"index": 342, "extra_field_not_allowed": true}
+{"index": 343, "extra_field_not_allowed": true}
+{"index": 344, "extra_field_not_allowed": true}
+{"index": 345, "extra_field_not_allowed": true}
+{"index": 346, "extra_field_not_allowed": true}
+{"index": 347, "extra_field_not_allowed": true}
+{"index": 348, "extra_field_not_allowed": true}
+{"index": 349, "extra_field_not_allowed": true}
+{"index": 350, "extra_field_not_allowed": true}
+{"index": 351, "extra_field_not_allowed": true}
+{"index": 352, "extra_field_not_allowed": true}
+{"index": 353, "extra_field_not_allowed": true}
+{"index": 354, "extra_field_not_allowed": true}
+{"index": 355, "extra_field_not_allowed": true}
+{"index": 356, "extra_field_not_allowed": true}
+{"index": 357, "extra_field_not_allowed": true}
+{"index": 358, "extra_field_not_allowed": true}
+{"index": 359, "extra_field_not_allowed": true}
+{"index": 360, "extra_field_not_allowed": true}
+{"index": 361, "extra_field_not_allowed": true}
+{"index": 362, "extra_field_not_allowed": true}
+{"index": 363, "extra_field_not_allowed": true}
+{"index": 364, "extra_field_not_allowed": true}
+{"index": 365, "extra_field_not_allowed": true}
+{"index": 366, "extra_field_not_allowed": true}
+{"index": 367, "extra_field_not_allowed": true}
+{"index": 368, "extra_field_not_allowed": true}
+{"index": 369, "extra_field_not_allowed": true}
+{"index": 370, "extra_field_not_allowed": true}
+{"index": 371, "extra_field_not_allowed": true}
+{"index": 372, "extra_field_not_allowed": true}
+{"index": 373, "extra_field_not_allowed": true}
+{"index": 374, "extra_field_not_allowed": true}
+{"index": 375, "extra_field_not_allowed": true}
+{"index": 376, "extra_field_not_allowed": true}
+{"index": 377, "extra_field_not_allowed": true}
+{"index": 378, "extra_field_not_allowed": true}
+{"index": 379, "extra_field_not_allowed": true}
+{"index": 380, "extra_field_not_allowed": true}
+{"index": 381, "extra_field_not_allowed": true}
+{"index": 382, "extra_field_not_allowed": true}
+{"index": 383, "extra_field_not_allowed": true}
+{"index": 384, "extra_field_not_allowed": true}
+{"index": 385, "extra_field_not_allowed": true}
+{"index": 386, "extra_field_not_allowed": true}
+{"index": 387, "extra_field_not_allowed": true}
+{"index": 388, "extra_field_not_allowed": true}
+{"index": 389, "extra_field_not_allowed": true}
+{"index": 390, "extra_field_not_allowed": true}
+{"index": 391, "extra_field_not_allowed": true}
+{"index": 392, "extra_field_not_allowed": true}
+{"index": 393, "extra_field_not_allowed": true}
+{"index": 394, "extra_field_not_allowed": true}
+{"index": 395, "extra_field_not_allowed": true}
+{"index": 396, "extra_field_not_allowed": true}
+{"index": 397, "extra_field_not_allowed": true}
+{"index": 398, "extra_field_not_allowed": true}
+{"index": 399, "extra_field_not_allowed": true}
+{"index": 400, "extra_field_not_allowed": true}
+{"index": 401, "extra_field_not_allowed": true}
+{"index": 402, "extra_field_not_allowed": true}
+{"index": 403, "extra_field_not_allowed": true}
+{"index": 404, "extra_field_not_allowed": true}
+{"index": 405, "extra_field_not_allowed": true}
+{"index": 406, "extra_field_not_allowed": true}
+{"index": 407, "extra_field_not_allowed": true}
+{"index": 408, "extra_field_not_allowed": true}
+{"index": 409, "extra_field_not_allowed": true}
+{"index": 410, "extra_field_not_allowed": true}
+{"index": 411, "extra_field_not_allowed": true}
+{"index": 412, "extra_field_not_allowed": true}
+{"index": 413, "extra_field_not_allowed": true}
+{"index": 414, "extra_field_not_allowed": true}
+{"index": 415, "extra_field_not_allowed": true}
+{"index": 416, "extra_field_not_allowed": true}
+{"index": 417, "extra_field_not_allowed": true}
+{"index": 418, "extra_field_not_allowed": true}
+{"index": 419, "extra_field_not_allowed": true}
+{"index": 420, "extra_field_not_allowed": true}
+{"index": 421, "extra_field_not_allowed": true}
+{"index": 422, "extra_field_not_allowed": true}
+{"index": 423, "extra_field_not_allowed": true}
+{"index": 424, "extra_field_not_allowed": true}
+{"index": 425, "extra_field_not_allowed": true}
+{"index": 426, "extra_field_not_allowed": true}
+{"index": 427, "extra_field_not_allowed": true}
+{"index": 428, "extra_field_not_allowed": true}
+{"index": 429, "extra_field_not_allowed": true}
+{"index": 430, "extra_field_not_allowed": true}
+{"index": 431, "extra_field_not_allowed": true}
+{"index": 432, "extra_field_not_allowed": true}
+{"index": 433, "extra_field_not_allowed": true}
+{"index": 434, "extra_field_not_allowed": true}
+{"index": 435, "extra_field_not_allowed": true}
+{"index": 436, "extra_field_not_allowed": true}
+{"index": 437, "extra_field_not_allowed": true}
+{"index": 438, "extra_field_not_allowed": true}
+{"index": 439, "extra_field_not_allowed": true}
+{"index": 440, "extra_field_not_allowed": true}
+{"index": 441, "extra_field_not_allowed": true}
+{"index": 442, "extra_field_not_allowed": true}
+{"index": 443, "extra_field_not_allowed": true}
+{"index": 444, "extra_field_not_allowed": true}
+{"index": 445, "extra_field_not_allowed": true}
+{"index": 446, "extra_field_not_allowed": true}
+{"index": 447, "extra_field_not_allowed": true}
+{"index": 448, "extra_field_not_allowed": true}
+{"index": 449, "extra_field_not_allowed": true}
+{"index": 450, "extra_field_not_allowed": true}
+{"index": 451, "extra_field_not_allowed": true}
+{"index": 452, "extra_field_not_allowed": true}
+{"index": 453, "extra_field_not_allowed": true}
+{"index": 454, "extra_field_not_allowed": true}
+{"index": 455, "extra_field_not_allowed": true}
+{"index": 456, "extra_field_not_allowed": true}
+{"index": 457, "extra_field_not_allowed": true}
+{"index": 458, "extra_field_not_allowed": true}
+{"index": 459, "extra_field_not_allowed": true}
+{"index": 460, "extra_field_not_allowed": true}
+{"index": 461, "extra_field_not_allowed": true}
+{"index": 462, "extra_field_not_allowed": true}
+{"index": 463, "extra_field_not_allowed": true}
+{"index": 464, "extra_field_not_allowed": true}
+{"index": 465, "extra_field_not_allowed": true}
+{"index": 466, "extra_field_not_allowed": true}
+{"index": 467, "extra_field_not_allowed": true}
+{"index": 468, "extra_field_not_allowed": true}
+{"index": 469, "extra_field_not_allowed": true}
+{"index": 470, "extra_field_not_allowed": true}
+{"index": 471, "extra_field_not_allowed": true}
+{"index": 472, "extra_field_not_allowed": true}
+{"index": 473, "extra_field_not_allowed": true}
+{"index": 474, "extra_field_not_allowed": true}
+{"index": 475, "extra_field_not_allowed": true}
+{"index": 476, "extra_field_not_allowed": true}
+{"index": 477, "extra_field_not_allowed": true}
+{"index": 478, "extra_field_not_allowed": true}
+{"index": 479, "extra_field_not_allowed": true}
+{"index": 480, "extra_field_not_allowed": true}
+{"index": 481, "extra_field_not_allowed": true}
+{"index": 482, "extra_field_not_allowed": true}
+{"index": 483, "extra_field_not_allowed": true}
+{"index": 484, "extra_field_not_allowed": true}
+{"index": 485, "extra_field_not_allowed": true}
+{"index": 486, "extra_field_not_allowed": true}
+{"index": 487, "extra_field_not_allowed": true}
+{"index": 488, "extra_field_not_allowed": true}
+{"index": 489, "extra_field_not_allowed": true}
+{"index": 490, "extra_field_not_allowed": true}
+{"index": 491, "extra_field_not_allowed": true}
+{"index": 492, "extra_field_not_allowed": true}
+{"index": 493, "extra_field_not_allowed": true}
+{"index": 494, "extra_field_not_allowed": true}
+{"index": 495, "extra_field_not_allowed": true}
+{"index": 496, "extra_field_not_allowed": true}
+{"index": 497, "extra_field_not_allowed": true}
+{"index": 498, "extra_field_not_allowed": true}
+{"index": 499, "extra_field_not_allowed": true}
+{"index": 500, "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/wrong_value.ndjson
new file mode 100644
index 000000000..16bceef8e
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-large/wrong_value.ndjson
@@ -0,0 +1,500 @@
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/canonical.ndjson
new file mode 100644
index 000000000..124467000
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/canonical.ndjson
@@ -0,0 +1,50 @@
+{"index": 1}
+{"index": 2}
+{"index": 3}
+{"index": 4}
+{"index": 5}
+{"index": 6}
+{"index": 7}
+{"index": 8}
+{"index": 9}
+{"index": 10}
+{"index": 11}
+{"index": 12}
+{"index": 13}
+{"index": 14}
+{"index": 15}
+{"index": 16}
+{"index": 17}
+{"index": 18}
+{"index": 19}
+{"index": 20}
+{"index": 21}
+{"index": 22}
+{"index": 23}
+{"index": 24}
+{"index": 25}
+{"index": 26}
+{"index": 27}
+{"index": 28}
+{"index": 29}
+{"index": 30}
+{"index": 31}
+{"index": 32}
+{"index": 33}
+{"index": 34}
+{"index": 35}
+{"index": 36}
+{"index": 37}
+{"index": 38}
+{"index": 39}
+{"index": 40}
+{"index": 41}
+{"index": 42}
+{"index": 43}
+{"index": 44}
+{"index": 45}
+{"index": 46}
+{"index": 47}
+{"index": 48}
+{"index": 49}
+{"index": 50}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/duplicate.ndjson
new file mode 100644
index 000000000..eacb8275f
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/duplicate.ndjson
@@ -0,0 +1,51 @@
+{"index": 1}
+{"index": 2}
+{"index": 3}
+{"index": 4}
+{"index": 5}
+{"index": 6}
+{"index": 7}
+{"index": 8}
+{"index": 9}
+{"index": 10}
+{"index": 11}
+{"index": 12}
+{"index": 13}
+{"index": 14}
+{"index": 15}
+{"index": 16}
+{"index": 17}
+{"index": 18}
+{"index": 19}
+{"index": 20}
+{"index": 21}
+{"index": 22}
+{"index": 23}
+{"index": 24}
+{"index": 25}
+{"index": 26}
+{"index": 27}
+{"index": 28}
+{"index": 29}
+{"index": 30}
+{"index": 31}
+{"index": 32}
+{"index": 33}
+{"index": 34}
+{"index": 35}
+{"index": 36}
+{"index": 37}
+{"index": 38}
+{"index": 39}
+{"index": 40}
+{"index": 41}
+{"index": 42}
+{"index": 43}
+{"index": 44}
+{"index": 45}
+{"index": 46}
+{"index": 47}
+{"index": 48}
+{"index": 49}
+{"index": 50}
+{"index": 50}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/extra.ndjson
new file mode 100644
index 000000000..426b4339a
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/extra.ndjson
@@ -0,0 +1,51 @@
+{"index": 1}
+{"index": 2}
+{"index": 3}
+{"index": 4}
+{"index": 5}
+{"index": 6}
+{"index": 7}
+{"index": 8}
+{"index": 9}
+{"index": 10}
+{"index": 11}
+{"index": 12}
+{"index": 13}
+{"index": 14}
+{"index": 15}
+{"index": 16}
+{"index": 17}
+{"index": 18}
+{"index": 19}
+{"index": 20}
+{"index": 21}
+{"index": 22}
+{"index": 23}
+{"index": 24}
+{"index": 25}
+{"index": 26}
+{"index": 27}
+{"index": 28}
+{"index": 29}
+{"index": 30}
+{"index": 31}
+{"index": 32}
+{"index": 33}
+{"index": 34}
+{"index": 35}
+{"index": 36}
+{"index": 37}
+{"index": 38}
+{"index": 39}
+{"index": 40}
+{"index": 41}
+{"index": 42}
+{"index": 43}
+{"index": 44}
+{"index": 45}
+{"index": 46}
+{"index": 47}
+{"index": 48}
+{"index": 49}
+{"index": 50}
+{"index": 1, "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/family_mut.ndjson
new file mode 100644
index 000000000..ae758564a
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/family_mut.ndjson
@@ -0,0 +1,50 @@
+{"index": 999}
+{"index": 2}
+{"index": 3}
+{"index": 4}
+{"index": 5}
+{"index": 6}
+{"index": 7}
+{"index": 8}
+{"index": 9}
+{"index": 10}
+{"index": 11}
+{"index": 12}
+{"index": 13}
+{"index": 14}
+{"index": 15}
+{"index": 16}
+{"index": 17}
+{"index": 18}
+{"index": 19}
+{"index": 20}
+{"index": 21}
+{"index": 22}
+{"index": 23}
+{"index": 24}
+{"index": 25}
+{"index": 26}
+{"index": 27}
+{"index": 28}
+{"index": 29}
+{"index": 30}
+{"index": 31}
+{"index": 32}
+{"index": 33}
+{"index": 34}
+{"index": 35}
+{"index": 36}
+{"index": 37}
+{"index": 38}
+{"index": 39}
+{"index": 40}
+{"index": 41}
+{"index": 42}
+{"index": 43}
+{"index": 44}
+{"index": 45}
+{"index": 46}
+{"index": 47}
+{"index": 48}
+{"index": 49}
+{"index": 50}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/missing.ndjson
new file mode 100644
index 000000000..a11506a99
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/missing.ndjson
@@ -0,0 +1,49 @@
+{"index": 1}
+{"index": 2}
+{"index": 3}
+{"index": 4}
+{"index": 5}
+{"index": 6}
+{"index": 7}
+{"index": 8}
+{"index": 9}
+{"index": 10}
+{"index": 11}
+{"index": 12}
+{"index": 13}
+{"index": 14}
+{"index": 15}
+{"index": 16}
+{"index": 17}
+{"index": 18}
+{"index": 19}
+{"index": 20}
+{"index": 21}
+{"index": 22}
+{"index": 23}
+{"index": 24}
+{"index": 25}
+{"index": 26}
+{"index": 27}
+{"index": 28}
+{"index": 29}
+{"index": 30}
+{"index": 31}
+{"index": 32}
+{"index": 33}
+{"index": 34}
+{"index": 35}
+{"index": 36}
+{"index": 37}
+{"index": 38}
+{"index": 39}
+{"index": 40}
+{"index": 41}
+{"index": 42}
+{"index": 43}
+{"index": 44}
+{"index": 45}
+{"index": 46}
+{"index": 47}
+{"index": 48}
+{"index": 49}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/wrong_field_set.ndjson
new file mode 100644
index 000000000..67b1f73a3
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/wrong_field_set.ndjson
@@ -0,0 +1,50 @@
+{"index": 1, "extra_field_not_allowed": true}
+{"index": 2, "extra_field_not_allowed": true}
+{"index": 3, "extra_field_not_allowed": true}
+{"index": 4, "extra_field_not_allowed": true}
+{"index": 5, "extra_field_not_allowed": true}
+{"index": 6, "extra_field_not_allowed": true}
+{"index": 7, "extra_field_not_allowed": true}
+{"index": 8, "extra_field_not_allowed": true}
+{"index": 9, "extra_field_not_allowed": true}
+{"index": 10, "extra_field_not_allowed": true}
+{"index": 11, "extra_field_not_allowed": true}
+{"index": 12, "extra_field_not_allowed": true}
+{"index": 13, "extra_field_not_allowed": true}
+{"index": 14, "extra_field_not_allowed": true}
+{"index": 15, "extra_field_not_allowed": true}
+{"index": 16, "extra_field_not_allowed": true}
+{"index": 17, "extra_field_not_allowed": true}
+{"index": 18, "extra_field_not_allowed": true}
+{"index": 19, "extra_field_not_allowed": true}
+{"index": 20, "extra_field_not_allowed": true}
+{"index": 21, "extra_field_not_allowed": true}
+{"index": 22, "extra_field_not_allowed": true}
+{"index": 23, "extra_field_not_allowed": true}
+{"index": 24, "extra_field_not_allowed": true}
+{"index": 25, "extra_field_not_allowed": true}
+{"index": 26, "extra_field_not_allowed": true}
+{"index": 27, "extra_field_not_allowed": true}
+{"index": 28, "extra_field_not_allowed": true}
+{"index": 29, "extra_field_not_allowed": true}
+{"index": 30, "extra_field_not_allowed": true}
+{"index": 31, "extra_field_not_allowed": true}
+{"index": 32, "extra_field_not_allowed": true}
+{"index": 33, "extra_field_not_allowed": true}
+{"index": 34, "extra_field_not_allowed": true}
+{"index": 35, "extra_field_not_allowed": true}
+{"index": 36, "extra_field_not_allowed": true}
+{"index": 37, "extra_field_not_allowed": true}
+{"index": 38, "extra_field_not_allowed": true}
+{"index": 39, "extra_field_not_allowed": true}
+{"index": 40, "extra_field_not_allowed": true}
+{"index": 41, "extra_field_not_allowed": true}
+{"index": 42, "extra_field_not_allowed": true}
+{"index": 43, "extra_field_not_allowed": true}
+{"index": 44, "extra_field_not_allowed": true}
+{"index": 45, "extra_field_not_allowed": true}
+{"index": 46, "extra_field_not_allowed": true}
+{"index": 47, "extra_field_not_allowed": true}
+{"index": 48, "extra_field_not_allowed": true}
+{"index": 49, "extra_field_not_allowed": true}
+{"index": 50, "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/wrong_value.ndjson
new file mode 100644
index 000000000..18b04fc6b
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/range-medium/wrong_value.ndjson
@@ -0,0 +1,50 @@
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
+{"index": 99999}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/canonical.ndjson
new file mode 100644
index 000000000..3820324e9
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/canonical.ndjson
@@ -0,0 +1,100 @@
+{"id": 1, "name": "User 1", "active": true}
+{"id": 2, "name": "User 2", "active": true}
+{"id": 3, "name": "User 3", "active": true}
+{"id": 4, "name": "User 4", "active": true}
+{"id": 5, "name": "User 5", "active": true}
+{"id": 6, "name": "User 6", "active": true}
+{"id": 7, "name": "User 7", "active": true}
+{"id": 8, "name": "User 8", "active": true}
+{"id": 9, "name": "User 9", "active": true}
+{"id": 10, "name": "User 10", "active": true}
+{"id": 11, "name": "User 11", "active": true}
+{"id": 12, "name": "User 12", "active": true}
+{"id": 13, "name": "User 13", "active": true}
+{"id": 14, "name": "User 14", "active": true}
+{"id": 15, "name": "User 15", "active": true}
+{"id": 16, "name": "User 16", "active": true}
+{"id": 17, "name": "User 17", "active": true}
+{"id": 18, "name": "User 18", "active": true}
+{"id": 19, "name": "User 19", "active": true}
+{"id": 20, "name": "User 20", "active": true}
+{"id": 21, "name": "User 21", "active": true}
+{"id": 22, "name": "User 22", "active": true}
+{"id": 23, "name": "User 23", "active": true}
+{"id": 24, "name": "User 24", "active": true}
+{"id": 25, "name": "User 25", "active": true}
+{"id": 26, "name": "User 26", "active": true}
+{"id": 27, "name": "User 27", "active": true}
+{"id": 28, "name": "User 28", "active": true}
+{"id": 29, "name": "User 29", "active": true}
+{"id": 30, "name": "User 30", "active": true}
+{"id": 31, "name": "User 31", "active": true}
+{"id": 32, "name": "User 32", "active": true}
+{"id": 33, "name": "User 33", "active": true}
+{"id": 34, "name": "User 34", "active": true}
+{"id": 35, "name": "User 35", "active": true}
+{"id": 36, "name": "User 36", "active": true}
+{"id": 37, "name": "User 37", "active": true}
+{"id": 38, "name": "User 38", "active": true}
+{"id": 39, "name": "User 39", "active": true}
+{"id": 40, "name": "User 40", "active": true}
+{"id": 41, "name": "User 41", "active": true}
+{"id": 42, "name": "User 42", "active": true}
+{"id": 43, "name": "User 43", "active": true}
+{"id": 44, "name": "User 44", "active": true}
+{"id": 45, "name": "User 45", "active": true}
+{"id": 46, "name": "User 46", "active": true}
+{"id": 47, "name": "User 47", "active": true}
+{"id": 48, "name": "User 48", "active": true}
+{"id": 49, "name": "User 49", "active": true}
+{"id": 50, "name": "User 50", "active": true}
+{"id": 51, "name": "User 51", "active": true}
+{"id": 52, "name": "User 52", "active": true}
+{"id": 53, "name": "User 53", "active": true}
+{"id": 54, "name": "User 54", "active": true}
+{"id": 55, "name": "User 55", "active": true}
+{"id": 56, "name": "User 56", "active": true}
+{"id": 57, "name": "User 57", "active": true}
+{"id": 58, "name": "User 58", "active": true}
+{"id": 59, "name": "User 59", "active": true}
+{"id": 60, "name": "User 60", "active": true}
+{"id": 61, "name": "User 61", "active": true}
+{"id": 62, "name": "User 62", "active": true}
+{"id": 63, "name": "User 63", "active": true}
+{"id": 64, "name": "User 64", "active": true}
+{"id": 65, "name": "User 65", "active": true}
+{"id": 66, "name": "User 66", "active": true}
+{"id": 67, "name": "User 67", "active": true}
+{"id": 68, "name": "User 68", "active": true}
+{"id": 69, "name": "User 69", "active": true}
+{"id": 70, "name": "User 70", "active": true}
+{"id": 71, "name": "User 71", "active": true}
+{"id": 72, "name": "User 72", "active": true}
+{"id": 73, "name": "User 73", "active": true}
+{"id": 74, "name": "User 74", "active": true}
+{"id": 75, "name": "User 75", "active": true}
+{"id": 76, "name": "User 76", "active": true}
+{"id": 77, "name": "User 77", "active": true}
+{"id": 78, "name": "User 78", "active": true}
+{"id": 79, "name": "User 79", "active": true}
+{"id": 80, "name": "User 80", "active": true}
+{"id": 81, "name": "User 81", "active": true}
+{"id": 82, "name": "User 82", "active": true}
+{"id": 83, "name": "User 83", "active": true}
+{"id": 84, "name": "User 84", "active": true}
+{"id": 85, "name": "User 85", "active": true}
+{"id": 86, "name": "User 86", "active": true}
+{"id": 87, "name": "User 87", "active": true}
+{"id": 88, "name": "User 88", "active": true}
+{"id": 89, "name": "User 89", "active": true}
+{"id": 90, "name": "User 90", "active": true}
+{"id": 91, "name": "User 91", "active": true}
+{"id": 92, "name": "User 92", "active": true}
+{"id": 93, "name": "User 93", "active": true}
+{"id": 94, "name": "User 94", "active": true}
+{"id": 95, "name": "User 95", "active": true}
+{"id": 96, "name": "User 96", "active": true}
+{"id": 97, "name": "User 97", "active": true}
+{"id": 98, "name": "User 98", "active": true}
+{"id": 99, "name": "User 99", "active": true}
+{"id": 100, "name": "User 100", "active": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/duplicate.ndjson
new file mode 100644
index 000000000..0140da1ce
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/duplicate.ndjson
@@ -0,0 +1,101 @@
+{"id": 1, "name": "User 1", "active": true}
+{"id": 2, "name": "User 2", "active": true}
+{"id": 3, "name": "User 3", "active": true}
+{"id": 4, "name": "User 4", "active": true}
+{"id": 5, "name": "User 5", "active": true}
+{"id": 6, "name": "User 6", "active": true}
+{"id": 7, "name": "User 7", "active": true}
+{"id": 8, "name": "User 8", "active": true}
+{"id": 9, "name": "User 9", "active": true}
+{"id": 10, "name": "User 10", "active": true}
+{"id": 11, "name": "User 11", "active": true}
+{"id": 12, "name": "User 12", "active": true}
+{"id": 13, "name": "User 13", "active": true}
+{"id": 14, "name": "User 14", "active": true}
+{"id": 15, "name": "User 15", "active": true}
+{"id": 16, "name": "User 16", "active": true}
+{"id": 17, "name": "User 17", "active": true}
+{"id": 18, "name": "User 18", "active": true}
+{"id": 19, "name": "User 19", "active": true}
+{"id": 20, "name": "User 20", "active": true}
+{"id": 21, "name": "User 21", "active": true}
+{"id": 22, "name": "User 22", "active": true}
+{"id": 23, "name": "User 23", "active": true}
+{"id": 24, "name": "User 24", "active": true}
+{"id": 25, "name": "User 25", "active": true}
+{"id": 26, "name": "User 26", "active": true}
+{"id": 27, "name": "User 27", "active": true}
+{"id": 28, "name": "User 28", "active": true}
+{"id": 29, "name": "User 29", "active": true}
+{"id": 30, "name": "User 30", "active": true}
+{"id": 31, "name": "User 31", "active": true}
+{"id": 32, "name": "User 32", "active": true}
+{"id": 33, "name": "User 33", "active": true}
+{"id": 34, "name": "User 34", "active": true}
+{"id": 35, "name": "User 35", "active": true}
+{"id": 36, "name": "User 36", "active": true}
+{"id": 37, "name": "User 37", "active": true}
+{"id": 38, "name": "User 38", "active": true}
+{"id": 39, "name": "User 39", "active": true}
+{"id": 40, "name": "User 40", "active": true}
+{"id": 41, "name": "User 41", "active": true}
+{"id": 42, "name": "User 42", "active": true}
+{"id": 43, "name": "User 43", "active": true}
+{"id": 44, "name": "User 44", "active": true}
+{"id": 45, "name": "User 45", "active": true}
+{"id": 46, "name": "User 46", "active": true}
+{"id": 47, "name": "User 47", "active": true}
+{"id": 48, "name": "User 48", "active": true}
+{"id": 49, "name": "User 49", "active": true}
+{"id": 50, "name": "User 50", "active": true}
+{"id": 51, "name": "User 51", "active": true}
+{"id": 52, "name": "User 52", "active": true}
+{"id": 53, "name": "User 53", "active": true}
+{"id": 54, "name": "User 54", "active": true}
+{"id": 55, "name": "User 55", "active": true}
+{"id": 56, "name": "User 56", "active": true}
+{"id": 57, "name": "User 57", "active": true}
+{"id": 58, "name": "User 58", "active": true}
+{"id": 59, "name": "User 59", "active": true}
+{"id": 60, "name": "User 60", "active": true}
+{"id": 61, "name": "User 61", "active": true}
+{"id": 62, "name": "User 62", "active": true}
+{"id": 63, "name": "User 63", "active": true}
+{"id": 64, "name": "User 64", "active": true}
+{"id": 65, "name": "User 65", "active": true}
+{"id": 66, "name": "User 66", "active": true}
+{"id": 67, "name": "User 67", "active": true}
+{"id": 68, "name": "User 68", "active": true}
+{"id": 69, "name": "User 69", "active": true}
+{"id": 70, "name": "User 70", "active": true}
+{"id": 71, "name": "User 71", "active": true}
+{"id": 72, "name": "User 72", "active": true}
+{"id": 73, "name": "User 73", "active": true}
+{"id": 74, "name": "User 74", "active": true}
+{"id": 75, "name": "User 75", "active": true}
+{"id": 76, "name": "User 76", "active": true}
+{"id": 77, "name": "User 77", "active": true}
+{"id": 78, "name": "User 78", "active": true}
+{"id": 79, "name": "User 79", "active": true}
+{"id": 80, "name": "User 80", "active": true}
+{"id": 81, "name": "User 81", "active": true}
+{"id": 82, "name": "User 82", "active": true}
+{"id": 83, "name": "User 83", "active": true}
+{"id": 84, "name": "User 84", "active": true}
+{"id": 85, "name": "User 85", "active": true}
+{"id": 86, "name": "User 86", "active": true}
+{"id": 87, "name": "User 87", "active": true}
+{"id": 88, "name": "User 88", "active": true}
+{"id": 89, "name": "User 89", "active": true}
+{"id": 90, "name": "User 90", "active": true}
+{"id": 91, "name": "User 91", "active": true}
+{"id": 92, "name": "User 92", "active": true}
+{"id": 93, "name": "User 93", "active": true}
+{"id": 94, "name": "User 94", "active": true}
+{"id": 95, "name": "User 95", "active": true}
+{"id": 96, "name": "User 96", "active": true}
+{"id": 97, "name": "User 97", "active": true}
+{"id": 98, "name": "User 98", "active": true}
+{"id": 99, "name": "User 99", "active": true}
+{"id": 100, "name": "User 100", "active": true}
+{"id": 100, "name": "User 100", "active": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/extra.ndjson
new file mode 100644
index 000000000..3e2010e0d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/extra.ndjson
@@ -0,0 +1,101 @@
+{"id": 1, "name": "User 1", "active": true}
+{"id": 2, "name": "User 2", "active": true}
+{"id": 3, "name": "User 3", "active": true}
+{"id": 4, "name": "User 4", "active": true}
+{"id": 5, "name": "User 5", "active": true}
+{"id": 6, "name": "User 6", "active": true}
+{"id": 7, "name": "User 7", "active": true}
+{"id": 8, "name": "User 8", "active": true}
+{"id": 9, "name": "User 9", "active": true}
+{"id": 10, "name": "User 10", "active": true}
+{"id": 11, "name": "User 11", "active": true}
+{"id": 12, "name": "User 12", "active": true}
+{"id": 13, "name": "User 13", "active": true}
+{"id": 14, "name": "User 14", "active": true}
+{"id": 15, "name": "User 15", "active": true}
+{"id": 16, "name": "User 16", "active": true}
+{"id": 17, "name": "User 17", "active": true}
+{"id": 18, "name": "User 18", "active": true}
+{"id": 19, "name": "User 19", "active": true}
+{"id": 20, "name": "User 20", "active": true}
+{"id": 21, "name": "User 21", "active": true}
+{"id": 22, "name": "User 22", "active": true}
+{"id": 23, "name": "User 23", "active": true}
+{"id": 24, "name": "User 24", "active": true}
+{"id": 25, "name": "User 25", "active": true}
+{"id": 26, "name": "User 26", "active": true}
+{"id": 27, "name": "User 27", "active": true}
+{"id": 28, "name": "User 28", "active": true}
+{"id": 29, "name": "User 29", "active": true}
+{"id": 30, "name": "User 30", "active": true}
+{"id": 31, "name": "User 31", "active": true}
+{"id": 32, "name": "User 32", "active": true}
+{"id": 33, "name": "User 33", "active": true}
+{"id": 34, "name": "User 34", "active": true}
+{"id": 35, "name": "User 35", "active": true}
+{"id": 36, "name": "User 36", "active": true}
+{"id": 37, "name": "User 37", "active": true}
+{"id": 38, "name": "User 38", "active": true}
+{"id": 39, "name": "User 39", "active": true}
+{"id": 40, "name": "User 40", "active": true}
+{"id": 41, "name": "User 41", "active": true}
+{"id": 42, "name": "User 42", "active": true}
+{"id": 43, "name": "User 43", "active": true}
+{"id": 44, "name": "User 44", "active": true}
+{"id": 45, "name": "User 45", "active": true}
+{"id": 46, "name": "User 46", "active": true}
+{"id": 47, "name": "User 47", "active": true}
+{"id": 48, "name": "User 48", "active": true}
+{"id": 49, "name": "User 49", "active": true}
+{"id": 50, "name": "User 50", "active": true}
+{"id": 51, "name": "User 51", "active": true}
+{"id": 52, "name": "User 52", "active": true}
+{"id": 53, "name": "User 53", "active": true}
+{"id": 54, "name": "User 54", "active": true}
+{"id": 55, "name": "User 55", "active": true}
+{"id": 56, "name": "User 56", "active": true}
+{"id": 57, "name": "User 57", "active": true}
+{"id": 58, "name": "User 58", "active": true}
+{"id": 59, "name": "User 59", "active": true}
+{"id": 60, "name": "User 60", "active": true}
+{"id": 61, "name": "User 61", "active": true}
+{"id": 62, "name": "User 62", "active": true}
+{"id": 63, "name": "User 63", "active": true}
+{"id": 64, "name": "User 64", "active": true}
+{"id": 65, "name": "User 65", "active": true}
+{"id": 66, "name": "User 66", "active": true}
+{"id": 67, "name": "User 67", "active": true}
+{"id": 68, "name": "User 68", "active": true}
+{"id": 69, "name": "User 69", "active": true}
+{"id": 70, "name": "User 70", "active": true}
+{"id": 71, "name": "User 71", "active": true}
+{"id": 72, "name": "User 72", "active": true}
+{"id": 73, "name": "User 73", "active": true}
+{"id": 74, "name": "User 74", "active": true}
+{"id": 75, "name": "User 75", "active": true}
+{"id": 76, "name": "User 76", "active": true}
+{"id": 77, "name": "User 77", "active": true}
+{"id": 78, "name": "User 78", "active": true}
+{"id": 79, "name": "User 79", "active": true}
+{"id": 80, "name": "User 80", "active": true}
+{"id": 81, "name": "User 81", "active": true}
+{"id": 82, "name": "User 82", "active": true}
+{"id": 83, "name": "User 83", "active": true}
+{"id": 84, "name": "User 84", "active": true}
+{"id": 85, "name": "User 85", "active": true}
+{"id": 86, "name": "User 86", "active": true}
+{"id": 87, "name": "User 87", "active": true}
+{"id": 88, "name": "User 88", "active": true}
+{"id": 89, "name": "User 89", "active": true}
+{"id": 90, "name": "User 90", "active": true}
+{"id": 91, "name": "User 91", "active": true}
+{"id": 92, "name": "User 92", "active": true}
+{"id": 93, "name": "User 93", "active": true}
+{"id": 94, "name": "User 94", "active": true}
+{"id": 95, "name": "User 95", "active": true}
+{"id": 96, "name": "User 96", "active": true}
+{"id": 97, "name": "User 97", "active": true}
+{"id": 98, "name": "User 98", "active": true}
+{"id": 99, "name": "User 99", "active": true}
+{"id": 100, "name": "User 100", "active": true}
+{"id": 1, "name": "User 1", "active": true, "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/family_mut.ndjson
new file mode 100644
index 000000000..8e2225066
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/family_mut.ndjson
@@ -0,0 +1,100 @@
+{"id": 1, "name": "User 100", "active": true}
+{"id": 2, "name": "User 2", "active": true}
+{"id": 3, "name": "User 3", "active": true}
+{"id": 4, "name": "User 4", "active": true}
+{"id": 5, "name": "User 5", "active": true}
+{"id": 6, "name": "User 6", "active": true}
+{"id": 7, "name": "User 7", "active": true}
+{"id": 8, "name": "User 8", "active": true}
+{"id": 9, "name": "User 9", "active": true}
+{"id": 10, "name": "User 10", "active": true}
+{"id": 11, "name": "User 11", "active": true}
+{"id": 12, "name": "User 12", "active": true}
+{"id": 13, "name": "User 13", "active": true}
+{"id": 14, "name": "User 14", "active": true}
+{"id": 15, "name": "User 15", "active": true}
+{"id": 16, "name": "User 16", "active": true}
+{"id": 17, "name": "User 17", "active": true}
+{"id": 18, "name": "User 18", "active": true}
+{"id": 19, "name": "User 19", "active": true}
+{"id": 20, "name": "User 20", "active": true}
+{"id": 21, "name": "User 21", "active": true}
+{"id": 22, "name": "User 22", "active": true}
+{"id": 23, "name": "User 23", "active": true}
+{"id": 24, "name": "User 24", "active": true}
+{"id": 25, "name": "User 25", "active": true}
+{"id": 26, "name": "User 26", "active": true}
+{"id": 27, "name": "User 27", "active": true}
+{"id": 28, "name": "User 28", "active": true}
+{"id": 29, "name": "User 29", "active": true}
+{"id": 30, "name": "User 30", "active": true}
+{"id": 31, "name": "User 31", "active": true}
+{"id": 32, "name": "User 32", "active": true}
+{"id": 33, "name": "User 33", "active": true}
+{"id": 34, "name": "User 34", "active": true}
+{"id": 35, "name": "User 35", "active": true}
+{"id": 36, "name": "User 36", "active": true}
+{"id": 37, "name": "User 37", "active": true}
+{"id": 38, "name": "User 38", "active": true}
+{"id": 39, "name": "User 39", "active": true}
+{"id": 40, "name": "User 40", "active": true}
+{"id": 41, "name": "User 41", "active": true}
+{"id": 42, "name": "User 42", "active": true}
+{"id": 43, "name": "User 43", "active": true}
+{"id": 44, "name": "User 44", "active": true}
+{"id": 45, "name": "User 45", "active": true}
+{"id": 46, "name": "User 46", "active": true}
+{"id": 47, "name": "User 47", "active": true}
+{"id": 48, "name": "User 48", "active": true}
+{"id": 49, "name": "User 49", "active": true}
+{"id": 50, "name": "User 50", "active": true}
+{"id": 51, "name": "User 51", "active": true}
+{"id": 52, "name": "User 52", "active": true}
+{"id": 53, "name": "User 53", "active": true}
+{"id": 54, "name": "User 54", "active": true}
+{"id": 55, "name": "User 55", "active": true}
+{"id": 56, "name": "User 56", "active": true}
+{"id": 57, "name": "User 57", "active": true}
+{"id": 58, "name": "User 58", "active": true}
+{"id": 59, "name": "User 59", "active": true}
+{"id": 60, "name": "User 60", "active": true}
+{"id": 61, "name": "User 61", "active": true}
+{"id": 62, "name": "User 62", "active": true}
+{"id": 63, "name": "User 63", "active": true}
+{"id": 64, "name": "User 64", "active": true}
+{"id": 65, "name": "User 65", "active": true}
+{"id": 66, "name": "User 66", "active": true}
+{"id": 67, "name": "User 67", "active": true}
+{"id": 68, "name": "User 68", "active": true}
+{"id": 69, "name": "User 69", "active": true}
+{"id": 70, "name": "User 70", "active": true}
+{"id": 71, "name": "User 71", "active": true}
+{"id": 72, "name": "User 72", "active": true}
+{"id": 73, "name": "User 73", "active": true}
+{"id": 74, "name": "User 74", "active": true}
+{"id": 75, "name": "User 75", "active": true}
+{"id": 76, "name": "User 76", "active": true}
+{"id": 77, "name": "User 77", "active": true}
+{"id": 78, "name": "User 78", "active": true}
+{"id": 79, "name": "User 79", "active": true}
+{"id": 80, "name": "User 80", "active": true}
+{"id": 81, "name": "User 81", "active": true}
+{"id": 82, "name": "User 82", "active": true}
+{"id": 83, "name": "User 83", "active": true}
+{"id": 84, "name": "User 84", "active": true}
+{"id": 85, "name": "User 85", "active": true}
+{"id": 86, "name": "User 86", "active": true}
+{"id": 87, "name": "User 87", "active": true}
+{"id": 88, "name": "User 88", "active": true}
+{"id": 89, "name": "User 89", "active": true}
+{"id": 90, "name": "User 90", "active": true}
+{"id": 91, "name": "User 91", "active": true}
+{"id": 92, "name": "User 92", "active": true}
+{"id": 93, "name": "User 93", "active": true}
+{"id": 94, "name": "User 94", "active": true}
+{"id": 95, "name": "User 95", "active": true}
+{"id": 96, "name": "User 96", "active": true}
+{"id": 97, "name": "User 97", "active": true}
+{"id": 98, "name": "User 98", "active": true}
+{"id": 99, "name": "User 99", "active": true}
+{"id": 100, "name": "User 100", "active": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/missing.ndjson
new file mode 100644
index 000000000..1e5937e33
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/missing.ndjson
@@ -0,0 +1,99 @@
+{"id": 1, "name": "User 1", "active": true}
+{"id": 2, "name": "User 2", "active": true}
+{"id": 3, "name": "User 3", "active": true}
+{"id": 4, "name": "User 4", "active": true}
+{"id": 5, "name": "User 5", "active": true}
+{"id": 6, "name": "User 6", "active": true}
+{"id": 7, "name": "User 7", "active": true}
+{"id": 8, "name": "User 8", "active": true}
+{"id": 9, "name": "User 9", "active": true}
+{"id": 10, "name": "User 10", "active": true}
+{"id": 11, "name": "User 11", "active": true}
+{"id": 12, "name": "User 12", "active": true}
+{"id": 13, "name": "User 13", "active": true}
+{"id": 14, "name": "User 14", "active": true}
+{"id": 15, "name": "User 15", "active": true}
+{"id": 16, "name": "User 16", "active": true}
+{"id": 17, "name": "User 17", "active": true}
+{"id": 18, "name": "User 18", "active": true}
+{"id": 19, "name": "User 19", "active": true}
+{"id": 20, "name": "User 20", "active": true}
+{"id": 21, "name": "User 21", "active": true}
+{"id": 22, "name": "User 22", "active": true}
+{"id": 23, "name": "User 23", "active": true}
+{"id": 24, "name": "User 24", "active": true}
+{"id": 25, "name": "User 25", "active": true}
+{"id": 26, "name": "User 26", "active": true}
+{"id": 27, "name": "User 27", "active": true}
+{"id": 28, "name": "User 28", "active": true}
+{"id": 29, "name": "User 29", "active": true}
+{"id": 30, "name": "User 30", "active": true}
+{"id": 31, "name": "User 31", "active": true}
+{"id": 32, "name": "User 32", "active": true}
+{"id": 33, "name": "User 33", "active": true}
+{"id": 34, "name": "User 34", "active": true}
+{"id": 35, "name": "User 35", "active": true}
+{"id": 36, "name": "User 36", "active": true}
+{"id": 37, "name": "User 37", "active": true}
+{"id": 38, "name": "User 38", "active": true}
+{"id": 39, "name": "User 39", "active": true}
+{"id": 40, "name": "User 40", "active": true}
+{"id": 41, "name": "User 41", "active": true}
+{"id": 42, "name": "User 42", "active": true}
+{"id": 43, "name": "User 43", "active": true}
+{"id": 44, "name": "User 44", "active": true}
+{"id": 45, "name": "User 45", "active": true}
+{"id": 46, "name": "User 46", "active": true}
+{"id": 47, "name": "User 47", "active": true}
+{"id": 48, "name": "User 48", "active": true}
+{"id": 49, "name": "User 49", "active": true}
+{"id": 50, "name": "User 50", "active": true}
+{"id": 51, "name": "User 51", "active": true}
+{"id": 52, "name": "User 52", "active": true}
+{"id": 53, "name": "User 53", "active": true}
+{"id": 54, "name": "User 54", "active": true}
+{"id": 55, "name": "User 55", "active": true}
+{"id": 56, "name": "User 56", "active": true}
+{"id": 57, "name": "User 57", "active": true}
+{"id": 58, "name": "User 58", "active": true}
+{"id": 59, "name": "User 59", "active": true}
+{"id": 60, "name": "User 60", "active": true}
+{"id": 61, "name": "User 61", "active": true}
+{"id": 62, "name": "User 62", "active": true}
+{"id": 63, "name": "User 63", "active": true}
+{"id": 64, "name": "User 64", "active": true}
+{"id": 65, "name": "User 65", "active": true}
+{"id": 66, "name": "User 66", "active": true}
+{"id": 67, "name": "User 67", "active": true}
+{"id": 68, "name": "User 68", "active": true}
+{"id": 69, "name": "User 69", "active": true}
+{"id": 70, "name": "User 70", "active": true}
+{"id": 71, "name": "User 71", "active": true}
+{"id": 72, "name": "User 72", "active": true}
+{"id": 73, "name": "User 73", "active": true}
+{"id": 74, "name": "User 74", "active": true}
+{"id": 75, "name": "User 75", "active": true}
+{"id": 76, "name": "User 76", "active": true}
+{"id": 77, "name": "User 77", "active": true}
+{"id": 78, "name": "User 78", "active": true}
+{"id": 79, "name": "User 79", "active": true}
+{"id": 80, "name": "User 80", "active": true}
+{"id": 81, "name": "User 81", "active": true}
+{"id": 82, "name": "User 82", "active": true}
+{"id": 83, "name": "User 83", "active": true}
+{"id": 84, "name": "User 84", "active": true}
+{"id": 85, "name": "User 85", "active": true}
+{"id": 86, "name": "User 86", "active": true}
+{"id": 87, "name": "User 87", "active": true}
+{"id": 88, "name": "User 88", "active": true}
+{"id": 89, "name": "User 89", "active": true}
+{"id": 90, "name": "User 90", "active": true}
+{"id": 91, "name": "User 91", "active": true}
+{"id": 92, "name": "User 92", "active": true}
+{"id": 93, "name": "User 93", "active": true}
+{"id": 94, "name": "User 94", "active": true}
+{"id": 95, "name": "User 95", "active": true}
+{"id": 96, "name": "User 96", "active": true}
+{"id": 97, "name": "User 97", "active": true}
+{"id": 98, "name": "User 98", "active": true}
+{"id": 99, "name": "User 99", "active": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/wrong_field_set.ndjson
new file mode 100644
index 000000000..38b858a47
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/wrong_field_set.ndjson
@@ -0,0 +1,100 @@
+{"name": "User 1", "active": true, "extra_field_not_allowed": true}
+{"name": "User 2", "active": true, "extra_field_not_allowed": true}
+{"name": "User 3", "active": true, "extra_field_not_allowed": true}
+{"name": "User 4", "active": true, "extra_field_not_allowed": true}
+{"name": "User 5", "active": true, "extra_field_not_allowed": true}
+{"name": "User 6", "active": true, "extra_field_not_allowed": true}
+{"name": "User 7", "active": true, "extra_field_not_allowed": true}
+{"name": "User 8", "active": true, "extra_field_not_allowed": true}
+{"name": "User 9", "active": true, "extra_field_not_allowed": true}
+{"name": "User 10", "active": true, "extra_field_not_allowed": true}
+{"name": "User 11", "active": true, "extra_field_not_allowed": true}
+{"name": "User 12", "active": true, "extra_field_not_allowed": true}
+{"name": "User 13", "active": true, "extra_field_not_allowed": true}
+{"name": "User 14", "active": true, "extra_field_not_allowed": true}
+{"name": "User 15", "active": true, "extra_field_not_allowed": true}
+{"name": "User 16", "active": true, "extra_field_not_allowed": true}
+{"name": "User 17", "active": true, "extra_field_not_allowed": true}
+{"name": "User 18", "active": true, "extra_field_not_allowed": true}
+{"name": "User 19", "active": true, "extra_field_not_allowed": true}
+{"name": "User 20", "active": true, "extra_field_not_allowed": true}
+{"name": "User 21", "active": true, "extra_field_not_allowed": true}
+{"name": "User 22", "active": true, "extra_field_not_allowed": true}
+{"name": "User 23", "active": true, "extra_field_not_allowed": true}
+{"name": "User 24", "active": true, "extra_field_not_allowed": true}
+{"name": "User 25", "active": true, "extra_field_not_allowed": true}
+{"name": "User 26", "active": true, "extra_field_not_allowed": true}
+{"name": "User 27", "active": true, "extra_field_not_allowed": true}
+{"name": "User 28", "active": true, "extra_field_not_allowed": true}
+{"name": "User 29", "active": true, "extra_field_not_allowed": true}
+{"name": "User 30", "active": true, "extra_field_not_allowed": true}
+{"name": "User 31", "active": true, "extra_field_not_allowed": true}
+{"name": "User 32", "active": true, "extra_field_not_allowed": true}
+{"name": "User 33", "active": true, "extra_field_not_allowed": true}
+{"name": "User 34", "active": true, "extra_field_not_allowed": true}
+{"name": "User 35", "active": true, "extra_field_not_allowed": true}
+{"name": "User 36", "active": true, "extra_field_not_allowed": true}
+{"name": "User 37", "active": true, "extra_field_not_allowed": true}
+{"name": "User 38", "active": true, "extra_field_not_allowed": true}
+{"name": "User 39", "active": true, "extra_field_not_allowed": true}
+{"name": "User 40", "active": true, "extra_field_not_allowed": true}
+{"name": "User 41", "active": true, "extra_field_not_allowed": true}
+{"name": "User 42", "active": true, "extra_field_not_allowed": true}
+{"name": "User 43", "active": true, "extra_field_not_allowed": true}
+{"name": "User 44", "active": true, "extra_field_not_allowed": true}
+{"name": "User 45", "active": true, "extra_field_not_allowed": true}
+{"name": "User 46", "active": true, "extra_field_not_allowed": true}
+{"name": "User 47", "active": true, "extra_field_not_allowed": true}
+{"name": "User 48", "active": true, "extra_field_not_allowed": true}
+{"name": "User 49", "active": true, "extra_field_not_allowed": true}
+{"name": "User 50", "active": true, "extra_field_not_allowed": true}
+{"name": "User 51", "active": true, "extra_field_not_allowed": true}
+{"name": "User 52", "active": true, "extra_field_not_allowed": true}
+{"name": "User 53", "active": true, "extra_field_not_allowed": true}
+{"name": "User 54", "active": true, "extra_field_not_allowed": true}
+{"name": "User 55", "active": true, "extra_field_not_allowed": true}
+{"name": "User 56", "active": true, "extra_field_not_allowed": true}
+{"name": "User 57", "active": true, "extra_field_not_allowed": true}
+{"name": "User 58", "active": true, "extra_field_not_allowed": true}
+{"name": "User 59", "active": true, "extra_field_not_allowed": true}
+{"name": "User 60", "active": true, "extra_field_not_allowed": true}
+{"name": "User 61", "active": true, "extra_field_not_allowed": true}
+{"name": "User 62", "active": true, "extra_field_not_allowed": true}
+{"name": "User 63", "active": true, "extra_field_not_allowed": true}
+{"name": "User 64", "active": true, "extra_field_not_allowed": true}
+{"name": "User 65", "active": true, "extra_field_not_allowed": true}
+{"name": "User 66", "active": true, "extra_field_not_allowed": true}
+{"name": "User 67", "active": true, "extra_field_not_allowed": true}
+{"name": "User 68", "active": true, "extra_field_not_allowed": true}
+{"name": "User 69", "active": true, "extra_field_not_allowed": true}
+{"name": "User 70", "active": true, "extra_field_not_allowed": true}
+{"name": "User 71", "active": true, "extra_field_not_allowed": true}
+{"name": "User 72", "active": true, "extra_field_not_allowed": true}
+{"name": "User 73", "active": true, "extra_field_not_allowed": true}
+{"name": "User 74", "active": true, "extra_field_not_allowed": true}
+{"name": "User 75", "active": true, "extra_field_not_allowed": true}
+{"name": "User 76", "active": true, "extra_field_not_allowed": true}
+{"name": "User 77", "active": true, "extra_field_not_allowed": true}
+{"name": "User 78", "active": true, "extra_field_not_allowed": true}
+{"name": "User 79", "active": true, "extra_field_not_allowed": true}
+{"name": "User 80", "active": true, "extra_field_not_allowed": true}
+{"name": "User 81", "active": true, "extra_field_not_allowed": true}
+{"name": "User 82", "active": true, "extra_field_not_allowed": true}
+{"name": "User 83", "active": true, "extra_field_not_allowed": true}
+{"name": "User 84", "active": true, "extra_field_not_allowed": true}
+{"name": "User 85", "active": true, "extra_field_not_allowed": true}
+{"name": "User 86", "active": true, "extra_field_not_allowed": true}
+{"name": "User 87", "active": true, "extra_field_not_allowed": true}
+{"name": "User 88", "active": true, "extra_field_not_allowed": true}
+{"name": "User 89", "active": true, "extra_field_not_allowed": true}
+{"name": "User 90", "active": true, "extra_field_not_allowed": true}
+{"name": "User 91", "active": true, "extra_field_not_allowed": true}
+{"name": "User 92", "active": true, "extra_field_not_allowed": true}
+{"name": "User 93", "active": true, "extra_field_not_allowed": true}
+{"name": "User 94", "active": true, "extra_field_not_allowed": true}
+{"name": "User 95", "active": true, "extra_field_not_allowed": true}
+{"name": "User 96", "active": true, "extra_field_not_allowed": true}
+{"name": "User 97", "active": true, "extra_field_not_allowed": true}
+{"name": "User 98", "active": true, "extra_field_not_allowed": true}
+{"name": "User 99", "active": true, "extra_field_not_allowed": true}
+{"name": "User 100", "active": true, "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/wrong_value.ndjson
new file mode 100644
index 000000000..baa9d940b
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-large/wrong_value.ndjson
@@ -0,0 +1,100 @@
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/canonical.ndjson
new file mode 100644
index 000000000..65e679e07
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/canonical.ndjson
@@ -0,0 +1,10 @@
+{"id": 1, "name": "User 1", "active": true}
+{"id": 2, "name": "User 2", "active": true}
+{"id": 3, "name": "User 3", "active": true}
+{"id": 4, "name": "User 4", "active": true}
+{"id": 5, "name": "User 5", "active": true}
+{"id": 6, "name": "User 6", "active": true}
+{"id": 7, "name": "User 7", "active": true}
+{"id": 8, "name": "User 8", "active": true}
+{"id": 9, "name": "User 9", "active": true}
+{"id": 10, "name": "User 10", "active": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/duplicate.ndjson
new file mode 100644
index 000000000..9101b038c
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/duplicate.ndjson
@@ -0,0 +1,11 @@
+{"id": 1, "name": "User 1", "active": true}
+{"id": 2, "name": "User 2", "active": true}
+{"id": 3, "name": "User 3", "active": true}
+{"id": 4, "name": "User 4", "active": true}
+{"id": 5, "name": "User 5", "active": true}
+{"id": 6, "name": "User 6", "active": true}
+{"id": 7, "name": "User 7", "active": true}
+{"id": 8, "name": "User 8", "active": true}
+{"id": 9, "name": "User 9", "active": true}
+{"id": 10, "name": "User 10", "active": true}
+{"id": 10, "name": "User 10", "active": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/extra.ndjson
new file mode 100644
index 000000000..9b6cf871f
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/extra.ndjson
@@ -0,0 +1,11 @@
+{"id": 1, "name": "User 1", "active": true}
+{"id": 2, "name": "User 2", "active": true}
+{"id": 3, "name": "User 3", "active": true}
+{"id": 4, "name": "User 4", "active": true}
+{"id": 5, "name": "User 5", "active": true}
+{"id": 6, "name": "User 6", "active": true}
+{"id": 7, "name": "User 7", "active": true}
+{"id": 8, "name": "User 8", "active": true}
+{"id": 9, "name": "User 9", "active": true}
+{"id": 10, "name": "User 10", "active": true}
+{"id": 1, "name": "User 1", "active": true, "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/family_mut.ndjson
new file mode 100644
index 000000000..1a82f4ac7
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/family_mut.ndjson
@@ -0,0 +1,10 @@
+{"id": 1, "name": "User 10", "active": true}
+{"id": 2, "name": "User 2", "active": true}
+{"id": 3, "name": "User 3", "active": true}
+{"id": 4, "name": "User 4", "active": true}
+{"id": 5, "name": "User 5", "active": true}
+{"id": 6, "name": "User 6", "active": true}
+{"id": 7, "name": "User 7", "active": true}
+{"id": 8, "name": "User 8", "active": true}
+{"id": 9, "name": "User 9", "active": true}
+{"id": 10, "name": "User 10", "active": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/missing.ndjson
new file mode 100644
index 000000000..471ef91b2
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/missing.ndjson
@@ -0,0 +1,9 @@
+{"id": 1, "name": "User 1", "active": true}
+{"id": 2, "name": "User 2", "active": true}
+{"id": 3, "name": "User 3", "active": true}
+{"id": 4, "name": "User 4", "active": true}
+{"id": 5, "name": "User 5", "active": true}
+{"id": 6, "name": "User 6", "active": true}
+{"id": 7, "name": "User 7", "active": true}
+{"id": 8, "name": "User 8", "active": true}
+{"id": 9, "name": "User 9", "active": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/wrong_field_set.ndjson
new file mode 100644
index 000000000..48e5b860a
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/wrong_field_set.ndjson
@@ -0,0 +1,10 @@
+{"name": "User 1", "active": true, "extra_field_not_allowed": true}
+{"name": "User 2", "active": true, "extra_field_not_allowed": true}
+{"name": "User 3", "active": true, "extra_field_not_allowed": true}
+{"name": "User 4", "active": true, "extra_field_not_allowed": true}
+{"name": "User 5", "active": true, "extra_field_not_allowed": true}
+{"name": "User 6", "active": true, "extra_field_not_allowed": true}
+{"name": "User 7", "active": true, "extra_field_not_allowed": true}
+{"name": "User 8", "active": true, "extra_field_not_allowed": true}
+{"name": "User 9", "active": true, "extra_field_not_allowed": true}
+{"name": "User 10", "active": true, "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/wrong_value.ndjson
new file mode 100644
index 000000000..0f05a5c55
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/rows-medium/wrong_value.ndjson
@@ -0,0 +1,10 @@
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
+{"id": 99999, "name": "WRONG", "active": false}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/canonical.ndjson
new file mode 100644
index 000000000..552dd5665
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/canonical.ndjson
@@ -0,0 +1,51 @@
+{"id": 1, "type": "A", "value": 100}
+{"id": 1, "type": "A", "value": 110}
+{"id": 1, "type": "A", "value": 120}
+{"id": 1, "type": "A", "value": 130}
+{"id": 1, "type": "A", "value": 140}
+{"id": 1, "type": "A", "value": 150}
+{"id": 1, "type": "A", "value": 160}
+{"id": 1, "type": "A", "value": 170}
+{"id": 1, "type": "A", "value": 180}
+{"id": 1, "type": "A", "value": 190}
+{"id": 1, "type": "A", "value": 200}
+{"id": 1, "type": "A", "value": 210}
+{"id": 1, "type": "A", "value": 220}
+{"id": 1, "type": "A", "value": 230}
+{"id": 1, "type": "A", "value": 240}
+{"id": 1, "type": "A", "value": 250}
+{"id": 1, "type": "A", "value": 260}
+{"id": 1, "type": "A", "value": 270}
+{"id": 1, "type": "A", "value": 280}
+{"id": 1, "type": "A", "value": 290}
+{"id": 1, "type": "A", "value": 300}
+{"id": 1, "type": "A", "value": 310}
+{"id": 1, "type": "A", "value": 320}
+{"id": 1, "type": "A", "value": 330}
+{"id": 1, "type": "A", "value": 340}
+{"id": 1, "type": "A", "value": 350}
+{"id": 1, "type": "A", "value": 360}
+{"id": 1, "type": "A", "value": 370}
+{"id": 1, "type": "A", "value": 380}
+{"id": 1, "type": "A", "value": 390}
+{"id": 1, "type": "A", "value": 400}
+{"id": 1, "type": "A", "value": 410}
+{"id": 1, "type": "A", "value": 420}
+{"id": 1, "type": "A", "value": 430}
+{"id": 1, "type": "A", "value": 440}
+{"id": 1, "type": "A", "value": 450}
+{"id": 1, "type": "A", "value": 460}
+{"id": 1, "type": "A", "value": 470}
+{"id": 1, "type": "A", "value": 480}
+{"id": 1, "type": "A", "value": 490}
+{"id": 1, "type": "A", "value": 500}
+{"id": 1, "type": "A", "value": 510}
+{"id": 1, "type": "A", "value": 520}
+{"id": 1, "type": "A", "value": 530}
+{"id": 1, "type": "A", "value": 540}
+{"id": 1, "type": "A", "value": 550}
+{"id": 1, "type": "A", "value": 560}
+{"id": 1, "type": "A", "value": 570}
+{"id": 1, "type": "A", "value": 580}
+{"id": 1, "type": "A", "value": 590}
+{"id": 1, "type": "A", "value": 600}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/duplicate.ndjson
new file mode 100644
index 000000000..d20ba71bf
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/duplicate.ndjson
@@ -0,0 +1,52 @@
+{"id": 1, "type": "A", "value": 100}
+{"id": 1, "type": "A", "value": 110}
+{"id": 1, "type": "A", "value": 120}
+{"id": 1, "type": "A", "value": 130}
+{"id": 1, "type": "A", "value": 140}
+{"id": 1, "type": "A", "value": 150}
+{"id": 1, "type": "A", "value": 160}
+{"id": 1, "type": "A", "value": 170}
+{"id": 1, "type": "A", "value": 180}
+{"id": 1, "type": "A", "value": 190}
+{"id": 1, "type": "A", "value": 200}
+{"id": 1, "type": "A", "value": 210}
+{"id": 1, "type": "A", "value": 220}
+{"id": 1, "type": "A", "value": 230}
+{"id": 1, "type": "A", "value": 240}
+{"id": 1, "type": "A", "value": 250}
+{"id": 1, "type": "A", "value": 260}
+{"id": 1, "type": "A", "value": 270}
+{"id": 1, "type": "A", "value": 280}
+{"id": 1, "type": "A", "value": 290}
+{"id": 1, "type": "A", "value": 300}
+{"id": 1, "type": "A", "value": 310}
+{"id": 1, "type": "A", "value": 320}
+{"id": 1, "type": "A", "value": 330}
+{"id": 1, "type": "A", "value": 340}
+{"id": 1, "type": "A", "value": 350}
+{"id": 1, "type": "A", "value": 360}
+{"id": 1, "type": "A", "value": 370}
+{"id": 1, "type": "A", "value": 380}
+{"id": 1, "type": "A", "value": 390}
+{"id": 1, "type": "A", "value": 400}
+{"id": 1, "type": "A", "value": 410}
+{"id": 1, "type": "A", "value": 420}
+{"id": 1, "type": "A", "value": 430}
+{"id": 1, "type": "A", "value": 440}
+{"id": 1, "type": "A", "value": 450}
+{"id": 1, "type": "A", "value": 460}
+{"id": 1, "type": "A", "value": 470}
+{"id": 1, "type": "A", "value": 480}
+{"id": 1, "type": "A", "value": 490}
+{"id": 1, "type": "A", "value": 500}
+{"id": 1, "type": "A", "value": 510}
+{"id": 1, "type": "A", "value": 520}
+{"id": 1, "type": "A", "value": 530}
+{"id": 1, "type": "A", "value": 540}
+{"id": 1, "type": "A", "value": 550}
+{"id": 1, "type": "A", "value": 560}
+{"id": 1, "type": "A", "value": 570}
+{"id": 1, "type": "A", "value": 580}
+{"id": 1, "type": "A", "value": 590}
+{"id": 1, "type": "A", "value": 600}
+{"id": 1, "type": "A", "value": 600}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/extra.ndjson
new file mode 100644
index 000000000..7697ef3a4
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/extra.ndjson
@@ -0,0 +1,52 @@
+{"id": 1, "type": "A", "value": 100}
+{"id": 1, "type": "A", "value": 110}
+{"id": 1, "type": "A", "value": 120}
+{"id": 1, "type": "A", "value": 130}
+{"id": 1, "type": "A", "value": 140}
+{"id": 1, "type": "A", "value": 150}
+{"id": 1, "type": "A", "value": 160}
+{"id": 1, "type": "A", "value": 170}
+{"id": 1, "type": "A", "value": 180}
+{"id": 1, "type": "A", "value": 190}
+{"id": 1, "type": "A", "value": 200}
+{"id": 1, "type": "A", "value": 210}
+{"id": 1, "type": "A", "value": 220}
+{"id": 1, "type": "A", "value": 230}
+{"id": 1, "type": "A", "value": 240}
+{"id": 1, "type": "A", "value": 250}
+{"id": 1, "type": "A", "value": 260}
+{"id": 1, "type": "A", "value": 270}
+{"id": 1, "type": "A", "value": 280}
+{"id": 1, "type": "A", "value": 290}
+{"id": 1, "type": "A", "value": 300}
+{"id": 1, "type": "A", "value": 310}
+{"id": 1, "type": "A", "value": 320}
+{"id": 1, "type": "A", "value": 330}
+{"id": 1, "type": "A", "value": 340}
+{"id": 1, "type": "A", "value": 350}
+{"id": 1, "type": "A", "value": 360}
+{"id": 1, "type": "A", "value": 370}
+{"id": 1, "type": "A", "value": 380}
+{"id": 1, "type": "A", "value": 390}
+{"id": 1, "type": "A", "value": 400}
+{"id": 1, "type": "A", "value": 410}
+{"id": 1, "type": "A", "value": 420}
+{"id": 1, "type": "A", "value": 430}
+{"id": 1, "type": "A", "value": 440}
+{"id": 1, "type": "A", "value": 450}
+{"id": 1, "type": "A", "value": 460}
+{"id": 1, "type": "A", "value": 470}
+{"id": 1, "type": "A", "value": 480}
+{"id": 1, "type": "A", "value": 490}
+{"id": 1, "type": "A", "value": 500}
+{"id": 1, "type": "A", "value": 510}
+{"id": 1, "type": "A", "value": 520}
+{"id": 1, "type": "A", "value": 530}
+{"id": 1, "type": "A", "value": 540}
+{"id": 1, "type": "A", "value": 550}
+{"id": 1, "type": "A", "value": 560}
+{"id": 1, "type": "A", "value": 570}
+{"id": 1, "type": "A", "value": 580}
+{"id": 1, "type": "A", "value": 590}
+{"id": 1, "type": "A", "value": 600}
+{"id": 1, "type": "A", "value": 100, "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/family_mut.ndjson
new file mode 100644
index 000000000..374e1221b
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/family_mut.ndjson
@@ -0,0 +1,51 @@
+{"id": 1, "type": "A", "value": 100}
+{"id": 1, "type": "A", "value": 110}
+{"id": 1, "type": "A", "value": 120}
+{"id": 1, "type": "A", "value": 130}
+{"id": 1, "type": "A", "value": 140}
+{"id": 1, "type": "A", "value": 150}
+{"id": 1, "type": "A", "value": 160}
+{"id": 1, "type": "A", "value": 170}
+{"id": 1, "type": "A", "value": 180}
+{"id": 1, "type": "A", "value": 190}
+{"id": 1, "type": "A", "value": 200}
+{"id": 1, "type": "A", "value": 210}
+{"id": 1, "type": "A", "value": 220}
+{"id": 1, "type": "A", "value": 230}
+{"id": 1, "type": "A", "value": 240}
+{"id": 1, "type": "A", "value": 250}
+{"id": 1, "type": "A", "value": 260}
+{"id": 1, "type": "A", "value": 270}
+{"id": 1, "type": "A", "value": 280}
+{"id": 1, "type": "A", "value": 290}
+{"id": 1, "type": "A", "value": 300}
+{"id": 1, "type": "A", "value": 310}
+{"id": 1, "type": "A", "value": 320}
+{"id": 1, "type": "A", "value": 330}
+{"id": 1, "type": "A", "value": 340}
+{"id": 1, "type": "A", "value": 350}
+{"id": 1, "type": "A", "value": 360}
+{"id": 1, "type": "A", "value": 370}
+{"id": 1, "type": "A", "value": 380}
+{"id": 1, "type": "A", "value": 390}
+{"id": 1, "type": "A", "value": 400}
+{"id": 1, "type": "A", "value": 410}
+{"id": 1, "type": "A", "value": 420}
+{"id": 1, "type": "A", "value": 430}
+{"id": 1, "type": "A", "value": 440}
+{"id": 1, "type": "A", "value": 450}
+{"id": 1, "type": "A", "value": 460}
+{"id": 1, "type": "A", "value": 470}
+{"id": 1, "type": "A", "value": 480}
+{"id": 1, "type": "A", "value": 490}
+{"id": 1, "type": "A", "value": 500}
+{"id": 1, "type": "A", "value": 510}
+{"id": 1, "type": "A", "value": 520}
+{"id": 1, "type": "A", "value": 530}
+{"id": 1, "type": "A", "value": 540}
+{"id": 1, "type": "A", "value": 550}
+{"id": 1, "type": "A", "value": 560}
+{"id": 1, "type": "A", "value": 570}
+{"id": 1, "type": "A", "value": 580}
+{"id": 1, "type": "A", "value": 590}
+{"id": 1, "type": "A", "value": 9999}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/missing.ndjson
new file mode 100644
index 000000000..0642302fc
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/missing.ndjson
@@ -0,0 +1,50 @@
+{"id": 1, "type": "A", "value": 100}
+{"id": 1, "type": "A", "value": 110}
+{"id": 1, "type": "A", "value": 120}
+{"id": 1, "type": "A", "value": 130}
+{"id": 1, "type": "A", "value": 140}
+{"id": 1, "type": "A", "value": 150}
+{"id": 1, "type": "A", "value": 160}
+{"id": 1, "type": "A", "value": 170}
+{"id": 1, "type": "A", "value": 180}
+{"id": 1, "type": "A", "value": 190}
+{"id": 1, "type": "A", "value": 200}
+{"id": 1, "type": "A", "value": 210}
+{"id": 1, "type": "A", "value": 220}
+{"id": 1, "type": "A", "value": 230}
+{"id": 1, "type": "A", "value": 240}
+{"id": 1, "type": "A", "value": 250}
+{"id": 1, "type": "A", "value": 260}
+{"id": 1, "type": "A", "value": 270}
+{"id": 1, "type": "A", "value": 280}
+{"id": 1, "type": "A", "value": 290}
+{"id": 1, "type": "A", "value": 300}
+{"id": 1, "type": "A", "value": 310}
+{"id": 1, "type": "A", "value": 320}
+{"id": 1, "type": "A", "value": 330}
+{"id": 1, "type": "A", "value": 340}
+{"id": 1, "type": "A", "value": 350}
+{"id": 1, "type": "A", "value": 360}
+{"id": 1, "type": "A", "value": 370}
+{"id": 1, "type": "A", "value": 380}
+{"id": 1, "type": "A", "value": 390}
+{"id": 1, "type": "A", "value": 400}
+{"id": 1, "type": "A", "value": 410}
+{"id": 1, "type": "A", "value": 420}
+{"id": 1, "type": "A", "value": 430}
+{"id": 1, "type": "A", "value": 440}
+{"id": 1, "type": "A", "value": 450}
+{"id": 1, "type": "A", "value": 460}
+{"id": 1, "type": "A", "value": 470}
+{"id": 1, "type": "A", "value": 480}
+{"id": 1, "type": "A", "value": 490}
+{"id": 1, "type": "A", "value": 500}
+{"id": 1, "type": "A", "value": 510}
+{"id": 1, "type": "A", "value": 520}
+{"id": 1, "type": "A", "value": 530}
+{"id": 1, "type": "A", "value": 540}
+{"id": 1, "type": "A", "value": 550}
+{"id": 1, "type": "A", "value": 560}
+{"id": 1, "type": "A", "value": 570}
+{"id": 1, "type": "A", "value": 580}
+{"id": 1, "type": "A", "value": 590}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/wrong_field_set.ndjson
new file mode 100644
index 000000000..aee2a6da4
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/wrong_field_set.ndjson
@@ -0,0 +1,51 @@
+{"type": "A", "value": 100, "extra_field_not_allowed": true}
+{"type": "A", "value": 110, "extra_field_not_allowed": true}
+{"type": "A", "value": 120, "extra_field_not_allowed": true}
+{"type": "A", "value": 130, "extra_field_not_allowed": true}
+{"type": "A", "value": 140, "extra_field_not_allowed": true}
+{"type": "A", "value": 150, "extra_field_not_allowed": true}
+{"type": "A", "value": 160, "extra_field_not_allowed": true}
+{"type": "A", "value": 170, "extra_field_not_allowed": true}
+{"type": "A", "value": 180, "extra_field_not_allowed": true}
+{"type": "A", "value": 190, "extra_field_not_allowed": true}
+{"type": "A", "value": 200, "extra_field_not_allowed": true}
+{"type": "A", "value": 210, "extra_field_not_allowed": true}
+{"type": "A", "value": 220, "extra_field_not_allowed": true}
+{"type": "A", "value": 230, "extra_field_not_allowed": true}
+{"type": "A", "value": 240, "extra_field_not_allowed": true}
+{"type": "A", "value": 250, "extra_field_not_allowed": true}
+{"type": "A", "value": 260, "extra_field_not_allowed": true}
+{"type": "A", "value": 270, "extra_field_not_allowed": true}
+{"type": "A", "value": 280, "extra_field_not_allowed": true}
+{"type": "A", "value": 290, "extra_field_not_allowed": true}
+{"type": "A", "value": 300, "extra_field_not_allowed": true}
+{"type": "A", "value": 310, "extra_field_not_allowed": true}
+{"type": "A", "value": 320, "extra_field_not_allowed": true}
+{"type": "A", "value": 330, "extra_field_not_allowed": true}
+{"type": "A", "value": 340, "extra_field_not_allowed": true}
+{"type": "A", "value": 350, "extra_field_not_allowed": true}
+{"type": "A", "value": 360, "extra_field_not_allowed": true}
+{"type": "A", "value": 370, "extra_field_not_allowed": true}
+{"type": "A", "value": 380, "extra_field_not_allowed": true}
+{"type": "A", "value": 390, "extra_field_not_allowed": true}
+{"type": "A", "value": 400, "extra_field_not_allowed": true}
+{"type": "A", "value": 410, "extra_field_not_allowed": true}
+{"type": "A", "value": 420, "extra_field_not_allowed": true}
+{"type": "A", "value": 430, "extra_field_not_allowed": true}
+{"type": "A", "value": 440, "extra_field_not_allowed": true}
+{"type": "A", "value": 450, "extra_field_not_allowed": true}
+{"type": "A", "value": 460, "extra_field_not_allowed": true}
+{"type": "A", "value": 470, "extra_field_not_allowed": true}
+{"type": "A", "value": 480, "extra_field_not_allowed": true}
+{"type": "A", "value": 490, "extra_field_not_allowed": true}
+{"type": "A", "value": 500, "extra_field_not_allowed": true}
+{"type": "A", "value": 510, "extra_field_not_allowed": true}
+{"type": "A", "value": 520, "extra_field_not_allowed": true}
+{"type": "A", "value": 530, "extra_field_not_allowed": true}
+{"type": "A", "value": 540, "extra_field_not_allowed": true}
+{"type": "A", "value": 550, "extra_field_not_allowed": true}
+{"type": "A", "value": 560, "extra_field_not_allowed": true}
+{"type": "A", "value": 570, "extra_field_not_allowed": true}
+{"type": "A", "value": 580, "extra_field_not_allowed": true}
+{"type": "A", "value": 590, "extra_field_not_allowed": true}
+{"type": "A", "value": 600, "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/wrong_value.ndjson
new file mode 100644
index 000000000..bc22b4ad0
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-large/wrong_value.ndjson
@@ -0,0 +1,51 @@
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/canonical.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/canonical.ndjson
new file mode 100644
index 000000000..60273c5b6
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/canonical.ndjson
@@ -0,0 +1,6 @@
+{"id": 1, "type": "A", "value": 100}
+{"id": 1, "type": "A", "value": 110}
+{"id": 1, "type": "A", "value": 120}
+{"id": 1, "type": "A", "value": 130}
+{"id": 1, "type": "A", "value": 140}
+{"id": 1, "type": "A", "value": 150}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/duplicate.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/duplicate.ndjson
new file mode 100644
index 000000000..5407ad3ff
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/duplicate.ndjson
@@ -0,0 +1,7 @@
+{"id": 1, "type": "A", "value": 100}
+{"id": 1, "type": "A", "value": 110}
+{"id": 1, "type": "A", "value": 120}
+{"id": 1, "type": "A", "value": 130}
+{"id": 1, "type": "A", "value": 140}
+{"id": 1, "type": "A", "value": 150}
+{"id": 1, "type": "A", "value": 150}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/empty.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/empty.ndjson
new file mode 100644
index 000000000..e69de29bb
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/extra.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/extra.ndjson
new file mode 100644
index 000000000..67414abe6
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/extra.ndjson
@@ -0,0 +1,7 @@
+{"id": 1, "type": "A", "value": 100}
+{"id": 1, "type": "A", "value": 110}
+{"id": 1, "type": "A", "value": 120}
+{"id": 1, "type": "A", "value": 130}
+{"id": 1, "type": "A", "value": 140}
+{"id": 1, "type": "A", "value": 150}
+{"id": 1, "type": "A", "value": 100, "extra_id": "fake"}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/family_mut.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/family_mut.ndjson
new file mode 100644
index 000000000..6046f3d4c
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/family_mut.ndjson
@@ -0,0 +1,6 @@
+{"id": 1, "type": "A", "value": 100}
+{"id": 1, "type": "A", "value": 110}
+{"id": 1, "type": "A", "value": 120}
+{"id": 1, "type": "A", "value": 130}
+{"id": 1, "type": "A", "value": 140}
+{"id": 1, "type": "A", "value": 9999}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/malformed.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/malformed.ndjson
new file mode 100644
index 000000000..8192bf86d
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/malformed.ndjson
@@ -0,0 +1 @@
+{"id": 1, "bad_json":
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/missing.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/missing.ndjson
new file mode 100644
index 000000000..fa3e1a829
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/missing.ndjson
@@ -0,0 +1,5 @@
+{"id": 1, "type": "A", "value": 100}
+{"id": 1, "type": "A", "value": 110}
+{"id": 1, "type": "A", "value": 120}
+{"id": 1, "type": "A", "value": 130}
+{"id": 1, "type": "A", "value": 140}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/wrong_field_set.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/wrong_field_set.ndjson
new file mode 100644
index 000000000..712d1f4c7
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/wrong_field_set.ndjson
@@ -0,0 +1,6 @@
+{"type": "A", "value": 100, "extra_field_not_allowed": true}
+{"type": "A", "value": 110, "extra_field_not_allowed": true}
+{"type": "A", "value": 120, "extra_field_not_allowed": true}
+{"type": "A", "value": 130, "extra_field_not_allowed": true}
+{"type": "A", "value": 140, "extra_field_not_allowed": true}
+{"type": "A", "value": 150, "extra_field_not_allowed": true}
diff --git a/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/wrong_value.ndjson b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/wrong_value.ndjson
new file mode 100644
index 000000000..68c229f49
--- /dev/null
+++ b/labs/20-deltawire/eval/analysis/testdata/synthetic/verifier_fixtures/variants-medium/wrong_value.ndjson
@@ -0,0 +1,6 @@
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
+{"id": 99999, "type": "WRONG", "value": 99999}
diff --git a/labs/20-deltawire/eval/docs/07-pr99-audit.md b/labs/20-deltawire/eval/docs/07-pr99-audit.md
new file mode 100644
index 000000000..cfe3e98df
--- /dev/null
+++ b/labs/20-deltawire/eval/docs/07-pr99-audit.md
@@ -0,0 +1,129 @@
+# PR #99 Mandatory Audit
+
+## 1. extract_runs.py manufactures rewards, tokens, hashes, correctness, and DeltaWire compliance.
+**Finding**: `extract_runs.py` hardcodes results instead of reading actual job artifacts.
+**Exact path**: `labs/20-deltawire/eval/analysis/extract_runs.py`
+**Exact lines or symbol**: `extracted.append({ ... "reward": 1.0, "semantic_oracle_result": {"exact_match": 1}, "input_tokens": 1000 ... })`
+**Reproduction command**: `cat labs/20-deltawire/eval/analysis/extract_runs.py`
+**Consequence**: Fabricates analysis results instead of performing real extraction.
+**Required correction**: Replace mock extraction with real parsing of Harbor artifacts.
+
+## 2. analyze_paired.py analyzes output tokens only.
+**Finding**: Reduces tokens by only calculating the difference in `output_tokens`.
+**Exact path**: `labs/20-deltawire/eval/analysis/analyze_paired.py`
+**Exact lines or symbol**: `b0_tokens = b0["output_tokens"]`
+**Reproduction command**: `grep "tokens" labs/20-deltawire/eval/analysis/analyze_paired.py`
+**Consequence**: Fails to compute `input_tokens + output_tokens` as required for total token reduction.
+**Required correction**: Calculate `total_tokens = input_tokens + output_tokens` and use that for reduction.
+
+## 3. geometric_mean_paired_reduction is assigned the arithmetic mean.
+**Finding**: Uses arithmetic mean instead of geometric mean.
+**Exact path**: `labs/20-deltawire/eval/analysis/analyze_paired.py`
+**Exact lines or symbol**: `analysis["geometric_mean_paired_reduction"] = analysis["aggregate_reduction"] # simplified`
+**Reproduction command**: `grep "geometric_mean" labs/20-deltawire/eval/analysis/analyze_paired.py`
+**Consequence**: Reports incorrect statistical data.
+**Required correction**: Implement correct geometric mean logic `1 - exp(mean(log(D1_i / B0_i)))`.
+
+## 4. task_clustered_95_interval is hardcoded to [0,0].
+**Finding**: Interval is hardcoded.
+**Exact path**: `labs/20-deltawire/eval/analysis/analyze_paired.py`
+**Exact lines or symbol**: `"task_clustered_95_interval": [0, 0],`
+**Reproduction command**: `grep "task_clustered_95_interval" labs/20-deltawire/eval/analysis/analyze_paired.py`
+**Consequence**: False reporting of confidence interval.
+**Required correction**: Implement actual fixed-seed task-clustered bootstrap interval.
+
+## 5. render_report.py hardcodes READY_FOR_72=true.
+**Finding**: Unconditionally reports true.
+**Exact path**: `labs/20-deltawire/eval/analysis/render_report.py`
+**Exact lines or symbol**: `## READY_FOR_72\ntrue`
+**Reproduction command**: `grep "READY_FOR_72" labs/20-deltawire/eval/analysis/render_report.py`
+**Consequence**: Bypasses readiness gate.
+**Required correction**: Read readiness from `readiness.json`.
+
+## 6. run_paired.sh remains a stub.
+**Finding**: `run_paired.sh` just prints a stub message and exits.
+**Exact path**: `labs/20-deltawire/eval/scripts/run_paired.sh`
+**Exact lines or symbol**: `echo 'run_paired.sh stub'`
+**Reproduction command**: `cat labs/20-deltawire/eval/scripts/run_paired.sh`
+**Consequence**: Does not actually execute any runs in Harbor.
+**Required correction**: Implement actual Harbor runner.
+
+## 7. run_canary.py and run_72_trials.py are referenced but not implemented.
+**Finding**: Missing scripts are recommended to users.
+**Exact path**: `labs/20-deltawire/eval/analysis/render_report.py`, `labs/20-deltawire/eval/scripts/generate_matrix.py`
+**Exact lines or symbol**: `python3 labs/20-deltawire/eval/scripts/run_72_trials.py`
+**Reproduction command**: `ls labs/20-deltawire/eval/scripts/run_canary.py labs/20-deltawire/eval/scripts/run_72_trials.py`
+**Consequence**: Users are instructed to run nonexistent commands.
+**Required correction**: Do not reference nonexistent scripts; implement CLI options in runner.
+
+## 8. frozen_72_manifest.json contains placeholder_sha256.
+**Finding**: Hardcoded placeholder strings for cryptographic evidence.
+**Exact path**: `labs/20-deltawire/eval/results/pilot-v0/frozen_72_manifest.json`
+**Exact lines or symbol**: `"deltawire_binary_sha256": "placeholder_sha256",`
+**Reproduction command**: `grep "placeholder" labs/20-deltawire/eval/results/pilot-v0/frozen_72_manifest.json`
+**Consequence**: Preflight is not cryptographically grounded.
+**Required correction**: Evaluate and include the actual binary SHA-256 hash.
+
+## 9. the manifest model was not grounded from the successful pilot.
+**Finding**: Arbitrary model choice (`gemini-2.5-pro`) was written.
+**Exact path**: `labs/20-deltawire/eval/results/pilot-v0/frozen_72_manifest.json`
+**Exact lines or symbol**: `"model": "gemini-2.5-pro"`
+**Reproduction command**: `grep "model" labs/20-deltawire/eval/results/pilot-v0/frozen_72_manifest.json`
+**Consequence**: Wrong model or placeholder model may be used instead of the verified pilot model.
+**Required correction**: Ground the exact model identifier from the actual pilot artifacts.
+
+## 10. per-run-results.jsonl, aggregate.json, provenance.json, and run-manifest.json are empty.
+**Finding**: Result files created by `touch` remain empty.
+**Exact path**: `labs/20-deltawire/eval/results/pilot-v0/`
+**Exact lines or symbol**: (Empty file)
+**Reproduction command**: `wc -c labs/20-deltawire/eval/results/pilot-v0/*.json*`
+**Consequence**: Provenance of 24-run pilot is missing.
+**Required correction**: Ground the exact source paths from the VM and extract the actual JSON properties into these files.
+
+## 11. semantic_oracle.py automatically accepts every task except three large tasks.
+**Finding**: Only 3 representative tasks check logic, everything else evaluates to `True`.
+**Exact path**: `labs/20-deltawire/eval/scripts/semantic_oracle.py`
+**Exact lines or symbol**: `else:\n        # Fallback for other tasks (we only canary 3 representatives)\n        passed = True`
+**Reproduction command**: `grep -A 2 "Fallback" labs/20-deltawire/eval/scripts/semantic_oracle.py`
+**Consequence**: Falsely passes output for 9 out of 12 tasks.
+**Required correction**: Implement full semantic validation logic for all 12 tasks independently.
+
+## 12. mixed-large and irregular-large do not enforce their claimed semantics.
+**Finding**: These functions check basic uniqueness and record counts but ignore task-specific schemas.
+**Exact path**: `labs/20-deltawire/eval/scripts/semantic_oracle.py`
+**Exact lines or symbol**: `def check_irregular_large(records):` and `def check_mixed_large(records):`
+**Reproduction command**: `cat labs/20-deltawire/eval/scripts/semantic_oracle.py`
+**Consequence**: Invalid output passes as long as it has enough random JSON objects.
+**Required correction**: Check the exact constraints required by their schemas.
+
+## 13. multiple task instructions contain only "# Instruction".
+**Finding**: Nine task instruction files are empty besides the header.
+**Exact path**: `labs/20-deltawire/eval/tasks/*/instruction.md`
+**Exact lines or symbol**: `# Instruction`
+**Reproduction command**: `cat labs/20-deltawire/eval/tasks/rows-large/instruction.md`
+**Consequence**: The agent has no instructions to generate data.
+**Required correction**: Write comprehensive and unambiguous instructions for all 12 tasks.
+
+## 14. test.sh selects the first NDJSON file rather than an exact output path.
+**Finding**: Verification target is arbitrarily the first NDJSON found in `testdata/generated/`.
+**Exact path**: `labs/20-deltawire/eval/tasks/*/tests/test.sh`
+**Exact lines or symbol**: `OUTPUT_FILE=$(ls testdata/generated/*.ndjson 2>/dev/null | head -n 1)`
+**Reproduction command**: `cat labs/20-deltawire/eval/tasks/matrix-large/tests/test.sh`
+**Consequence**: Ambiguous verification if multiple files exist; fails to enforce exact path requirement.
+**Required correction**: Hardcode the exact required output path.
+
+## 15. the task environments do not yet prove that Python and the oracle script are available at verifier runtime.
+**Finding**: `test.sh` relies on `../../scripts/semantic_oracle.py`.
+**Exact path**: `labs/20-deltawire/eval/tasks/*/tests/test.sh`
+**Exact lines or symbol**: `python3 ../../scripts/semantic_oracle.py`
+**Reproduction command**: `grep python3 labs/20-deltawire/eval/tasks/*/tests/test.sh`
+**Consequence**: Task will likely fail in the isolated Harbor task environment due to missing dependencies and incorrect relative paths.
+**Required correction**: Explicitly mount the script or copy it during test execution and ensure Python is present.
+
+## 16. the generated six-run report is based entirely on synthetic rows.
+**Finding**: The report output comes directly from the hardcoded `extract_runs.py` mock objects.
+**Exact path**: `labs/20-deltawire/eval/analysis/extract_runs.py`
+**Exact lines or symbol**: `extracted.append({ ... })`
+**Reproduction command**: `cat labs/20-deltawire/eval/analysis/extract_runs.py`
+**Consequence**: The report provides a fake appearance of having executed a canary test.
+**Required correction**: Move synthetic data generation to tests, enforce a validation gate against mock terminology, and implement actual execution analysis.
diff --git a/labs/20-deltawire/eval/manifests/frozen-72-manifest.json b/labs/20-deltawire/eval/manifests/frozen-72-manifest.json
new file mode 100644
index 000000000..8dc8aa01f
--- /dev/null
+++ b/labs/20-deltawire/eval/manifests/frozen-72-manifest.json
@@ -0,0 +1,90 @@
+{
+  "schema_version": "deltawire-experiment-manifest.v2",
+  "repository_commit": "04d9cd021be8d81cd2fa11eed6c41108011c6683",
+  "deltawire_binary_sha256": "453819efd1afe30b7e1815507de67b3150374955970a77231ce89a7a738c33c6",
+  "model": "google/gemini-3.1-pro-preview",
+  "expected_reported_model": "gemini-3.1-pro-preview",
+  "agent": "gemini-cli",
+  "harbor_package": "harbor==0.20.0",
+  "harbor_version": "0.20.0",
+  "harbor_command": [
+    "uvx",
+    "--from",
+    "harbor==0.20.0",
+    "harbor"
+  ],
+  "exact_tasks": [
+    "irregular-large",
+    "irregular-medium",
+    "matrix-large",
+    "matrix-medium",
+    "mixed-large",
+    "mixed-medium",
+    "range-large",
+    "range-medium",
+    "rows-large",
+    "rows-medium",
+    "variants-large",
+    "variants-medium"
+  ],
+  "exact_arms": [
+    "B0",
+    "D1"
+  ],
+  "exact_repetitions": 3,
+  "exact_expected_run_count": 72,
+  "matrix_seed": 42,
+  "input_hashes": {
+    "labs/20-deltawire/eval/skills/deltawire/SKILL.md": "1fa7d2547509a1fa937be37c87b60e3dce6ae72e2d282e0f35ee868f8da947d3",
+    "labs/20-deltawire/eval/scripts/runner.py": "a1fc22a99f197d3e75676f995f4b8fd64e1cc5696e342d2456e7b115ee615832",
+    "labs/20-deltawire/eval/scripts/fidelity.py": "f4779fdff12bea157d6fa29c0c296273a1644d6ab8384788d5abfec0344539cb",
+    "labs/20-deltawire/eval/scripts/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/analysis/extract_runs.py": "e00b16d693634421f18347decea7848a0262410f0b62cfc3ef681310e616f6ec",
+    "labs/20-deltawire/eval/analysis/analyze_paired.py": "0ade5cfefc36470b50878c03ec5363fb2ec74096ba68eb153598dc6f2e32dc96",
+    "labs/20-deltawire/eval/matrices/72-run-matrix.json": "f9dcddc4b7336418c7e2b8e7a7e51ca1e42b5a137255ff0823a732618f3568d5",
+    "labs/20-deltawire/eval/tasks/irregular-large/task-spec.json": "eaba8c76908cd4763564c39b8fc7814737d0a778a51b5639b5f53ad84159f29b",
+    "labs/20-deltawire/eval/tasks/irregular-medium/task-spec.json": "6d9fb29a5a96bcad6b8dd32cdb79505c2568919810e7273e51776d1e23ec6154",
+    "labs/20-deltawire/eval/tasks/matrix-large/task-spec.json": "c3b915e22e66a84068154f7e3c1415c8da3f1f2bc27c58f95625491719ecd37a",
+    "labs/20-deltawire/eval/tasks/matrix-medium/task-spec.json": "662d0f64e14960227082bbc7945a5545875f4bd58f543823ca8200141047bdaf",
+    "labs/20-deltawire/eval/tasks/mixed-large/task-spec.json": "27325165edd96839b32538f0bea956b2d23c0eb3030cf64f4820d878c680268c",
+    "labs/20-deltawire/eval/tasks/mixed-medium/task-spec.json": "f9162a958f49ff78817b5775401156d458de468093bd0ff5fabfbf8ac97205d0",
+    "labs/20-deltawire/eval/tasks/range-large/task-spec.json": "3ed07b27866c4a9d6fd33bb39239452e4f84c0809e7055c5353be601ef0c2130",
+    "labs/20-deltawire/eval/tasks/range-medium/task-spec.json": "3e4698b14d5c336db999d4ef9a070b796602f9048a6df2132bc41e1a36268f52",
+    "labs/20-deltawire/eval/tasks/rows-large/task-spec.json": "4290189b00ad70718388bd21a30aadb99a3097876dfd47500b252e23d7af6e39",
+    "labs/20-deltawire/eval/tasks/rows-medium/task-spec.json": "de0cf296ad54df2b2caebce20f03bf9a62136f2276cbd7622cf0740b9287ada8",
+    "labs/20-deltawire/eval/tasks/variants-large/task-spec.json": "7f8a251cd2b727b537c686103b7f825d6436e3b5a80f8768269ba186f5c1e1d3",
+    "labs/20-deltawire/eval/tasks/variants-medium/task-spec.json": "624959ec6b13d27c7a60cfcb35a1b8b0ab64584166fcd8164ecaefd0c0d2ff3d",
+    "labs/20-deltawire/eval/tasks/irregular-large/instruction.md": "430a3add507430f9ae996869abe58d87fb522b2bd3fb66b83dc283a22af282b7",
+    "labs/20-deltawire/eval/tasks/irregular-medium/instruction.md": "4e5943d4abf9abfbb1c0ba58d314cb097ae3d47c1a61ff846c7ab9d3cff85f37",
+    "labs/20-deltawire/eval/tasks/matrix-large/instruction.md": "cd517babe98628de00db3abdbf1ad0d844c02e335bd86dfe7d630f43ac9926f1",
+    "labs/20-deltawire/eval/tasks/matrix-medium/instruction.md": "9bdb909955989c16b99a701516d846e786366236de3eec5c11530c4e86d7494a",
+    "labs/20-deltawire/eval/tasks/mixed-large/instruction.md": "7f50eceb87f2fa514a3ac57f1f88b9c700837bc615496b7361db9e368986299d",
+    "labs/20-deltawire/eval/tasks/mixed-medium/instruction.md": "a3c69bc1c248ef141bb52f1bda27b62990f7175cb4b0728f8ec74210eddc4f4a",
+    "labs/20-deltawire/eval/tasks/range-large/instruction.md": "a234ea4d4045a1c29f6c6e10ddf0234873cf02fd2693bec1ae080bcded67c5f9",
+    "labs/20-deltawire/eval/tasks/range-medium/instruction.md": "a4d21822d0dfd9eddd9346a5aee8a4a6681569c6279c4934bf8301ad91844e7b",
+    "labs/20-deltawire/eval/tasks/rows-large/instruction.md": "c54b1b48d0a234238b06e34491403a3f946a991e27cbbb2b7c1a7cbbc17ec24e",
+    "labs/20-deltawire/eval/tasks/rows-medium/instruction.md": "8398ae4c2bfbe384d7f566ff65c02313261ac26ecfaac38d6521d00b782e56cf",
+    "labs/20-deltawire/eval/tasks/variants-large/instruction.md": "b4c243fe51bd4a8b15d2193307c43b5048b5e00c33baca730a3dd4a259a1bc8a",
+    "labs/20-deltawire/eval/tasks/variants-medium/instruction.md": "cd41f7bedefcf02ba839057e64ed6ff978771684d99fdccf0045a30d6aef00a5",
+    "labs/20-deltawire/eval/tasks/irregular-large/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/irregular-medium/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/matrix-large/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/matrix-medium/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/mixed-large/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/mixed-medium/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/range-large/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/range-medium/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/rows-large/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/rows-medium/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/variants-large/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557",
+    "labs/20-deltawire/eval/tasks/variants-medium/environment/Dockerfile": "3ca83b3aa2022348aa975fb3de7ec2fd540bc586426afbb65b051d907d9c1557"
+  },
+  "token_field_definitions": {
+    "input_tokens": "/agent_result/n_input_tokens",
+    "output_tokens": "/agent_result/n_output_tokens",
+    "cache_tokens": "/agent_result/n_cache_tokens"
+  },
+  "correctness_definition": "Contract-driven semantic oracle exact_match=1.",
+  "failure_and_rerun_policy": "Stop on infrastructure or fidelity failure; no automatic reruns.",
+  "analysis_version": "v2.0.0"
+}
diff --git a/labs/20-deltawire/eval/manifests/preflight-v2-evidence-lock.json b/labs/20-deltawire/eval/manifests/preflight-v2-evidence-lock.json
new file mode 100644
index 000000000..29a0afef6
--- /dev/null
+++ b/labs/20-deltawire/eval/manifests/preflight-v2-evidence-lock.json
@@ -0,0 +1,17 @@
+{
+  "evidence": "range-large-v2-failed-treatment-probe",
+  "immutable_files": {
+    "labs/20-deltawire/eval/manifests/preflight-v2-range-large.json": "f2cf575109b87ceca5255fdd700a1e6d913c48582a60ab8daa87523446544fc2",
+    "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/D1-plan-contract-receipt.json": "9dfe8eccf671b0ad67f14dfaa5340ec00568ee0fe4b26e56b4424af00672a00b",
+    "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/D1-treatment-fidelity.json": "4baa77436206f2e1e9148ad33135568345d54f9cead50af2abeb91ca65507ab3",
+    "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/range-large-v2__qCX5xEj/agent/gemini-cli.trajectory.jsonl": "f075b4ca6cf3352e5704d992f1ebcd09f50d3a42853675b0ab2ec82dc7977403",
+    "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/range-large-v2__qCX5xEj/artifacts/.deltawire/plan-contract-receipt.json": "e59511ff336e46b891259d12df37eaeb77c23761c37ed47d74330c08df8b38d1",
+    "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/range-large-v2__qCX5xEj/artifacts/.deltawire/plans/range-large.dw.json": "2d0b13459d191e62a02876b3404ae657037a4c0f853d3eb795140aed48e219dc",
+    "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/range-large-v2__qCX5xEj/artifacts/testdata/generated/output.ndjson": "9d205b57ab166d23c81927899f99b6ccece261c679c96979c58fbde975637776",
+    "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/range-large-v2__qCX5xEj/result.json": "4afb5cfaa8bf224a70ce82e4b152f2ff9b360ff8ac12079bd3b535fff71b149e",
+    "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/result.json": "6fb00cdf5e02819d48e50ab83a1c9fce5be0bc19bbf49777dc463252a3f6ff88",
+    "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/report.md": "ad327f5568500b0cc333827e49ac13b6269f813cd3a187c9a20eb376813bdfcf",
+    "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/run-ledger.json": "01d9afa06b084725d53afd9bb98121afaa8bf5fd91ba807ea7f3117a5aad6c22"
+  },
+  "schema_version": "evidence-lock.v1"
+}
diff --git a/labs/20-deltawire/eval/manifests/preflight-v2-range-large.json b/labs/20-deltawire/eval/manifests/preflight-v2-range-large.json
new file mode 100644
index 000000000..fc564530b
--- /dev/null
+++ b/labs/20-deltawire/eval/manifests/preflight-v2-range-large.json
@@ -0,0 +1,53 @@
+{
+  "agent": "gemini-cli",
+  "deltawire_binary_sha256": "453819efd1afe30b7e1815507de67b3150374955970a77231ce89a7a738c33c6",
+  "expected_reported_model": "gemini-3.1-pro-preview",
+  "failure_policy": "Stop on any failed gate; never rerun v1 or launch the 72-run matrix.",
+  "harbor_command": [
+    "uvx",
+    "--from",
+    "harbor==0.20.0",
+    "harbor"
+  ],
+  "harbor_package": "harbor==0.20.0",
+  "harbor_version": "0.20.0",
+  "input_hashes": {
+    "labs/20-deltawire/eval/matrices/72-run-matrix.json": "f9dcddc4b7336418c7e2b8e7a7e51ca1e42b5a137255ff0823a732618f3568d5",
+    "labs/20-deltawire/eval/probes/range-large-v2/authoritative-schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/probes/range-large-v2/environment/.deltawire/config.json": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+    "labs/20-deltawire/eval/probes/range-large-v2/environment/.deltawire/schemas/range-large.schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/probes/range-large-v2/environment/Dockerfile": "0d4281bcb20ed360a4815c6170b1bd3fa65beaf51942d7203809b075ca4d3ead",
+    "labs/20-deltawire/eval/probes/range-large-v2/environment/task-contract.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v2/environment/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/probes/range-large-v2/instruction.md": "b88ebb9274a2f6522eda2ca65c6ff257fa31810400307f07a3e28bfc5517a9bf",
+    "labs/20-deltawire/eval/probes/range-large-v2/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v2/task.toml": "e73667ceb433a0f6b0e016366d78118f44351b5942fdb01dab6be8ae5630f88a",
+    "labs/20-deltawire/eval/probes/range-large-v2/tests/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/probes/range-large-v2/tests/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v2/tests/test.sh": "8fc106530f63500be0aace428a5dbf3e08567feb2adc40968029cb7d7a95afbc",
+    "labs/20-deltawire/eval/scripts/fidelity.py": "3f294ff1dc53367076b5736fa0cd6036082792cb3dc6cc193e288db01f4d4421",
+    "labs/20-deltawire/eval/scripts/runner.py": "74c292f1e352f1fae344d51e01a7f2297d58e965b7e57a9983b82e545188b20e",
+    "labs/20-deltawire/eval/scripts/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/scripts/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/skills/deltawire/SKILL.md": "ba0f9c585b744c9974796bdff58bd07a6a25d5138ffd6ef807b4611caeee6daf",
+    "labs/20-deltawire/eval/tasks/range-large/authoritative-schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/config.json": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+    "labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/schemas/range-large.schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/tasks/range-large/environment/Dockerfile": "0d4281bcb20ed360a4815c6170b1bd3fa65beaf51942d7203809b075ca4d3ead",
+    "labs/20-deltawire/eval/tasks/range-large/environment/task-contract.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/tasks/range-large/environment/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/tasks/range-large/instruction.md": "549f4047b1a4ba4f7a3f3dce9e60d1d70b141caa189aafdba942b15e6d022aee",
+    "labs/20-deltawire/eval/tasks/range-large/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/tasks/range-large/task.toml": "b2523e7a94151afa1b765d9e42399cbe1afe6bc54129e223799e5443a1521b15",
+    "labs/20-deltawire/eval/tasks/range-large/tests/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/tasks/range-large/tests/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/tasks/range-large/tests/test.sh": "8fc106530f63500be0aace428a5dbf3e08567feb2adc40968029cb7d7a95afbc"
+  },
+  "model": "google/gemini-3.1-pro-preview",
+  "repository_commit": "ea6ae5d8",
+  "schema_version": "deltawire-preflight-v2-manifest.v1",
+  "scope": [
+    "probe/range-large-v2",
+    "range-large/r1"
+  ]
+}
diff --git a/labs/20-deltawire/eval/manifests/preflight-v3-evidence-lock.json b/labs/20-deltawire/eval/manifests/preflight-v3-evidence-lock.json
new file mode 100644
index 000000000..df4534a91
--- /dev/null
+++ b/labs/20-deltawire/eval/manifests/preflight-v3-evidence-lock.json
@@ -0,0 +1,72 @@
+{
+  "immutable_files": {
+    "labs/20-deltawire/eval/manifests/preflight-v2-evidence-lock.json": "dfaa4350681a36bcadef1e00d52bd90e16b2cf72638928ba2d0bc42b7af8f1b8",
+    "labs/20-deltawire/eval/manifests/preflight-v3-range-large.json": "bc219bbabbc8765a036b873313bed7aed62ffdcee0da570c703b64130d278a29",
+    "labs/20-deltawire/eval/matrices/72-run-matrix.json": "f9dcddc4b7336418c7e2b8e7a7e51ca1e42b5a137255ff0823a732618f3568d5",
+    "labs/20-deltawire/eval/probes/range-large-v3/authoritative-schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/probes/range-large-v3/environment/.deltawire/config.json": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+    "labs/20-deltawire/eval/probes/range-large-v3/environment/.deltawire/schemas/range-large.schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/probes/range-large-v3/environment/Dockerfile": "0d4281bcb20ed360a4815c6170b1bd3fa65beaf51942d7203809b075ca4d3ead",
+    "labs/20-deltawire/eval/probes/range-large-v3/environment/task-contract.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v3/environment/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/probes/range-large-v3/instruction.md": "2c1086bd55056b63c89bbb359fdb7a32168ce6c9598d02d6c084a52ad46e9af9",
+    "labs/20-deltawire/eval/probes/range-large-v3/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v3/task.toml": "854a1379edee3d090dfe85265753bcd045d192ccbd21e71a2003ac037436a49a",
+    "labs/20-deltawire/eval/probes/range-large-v3/tests/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/probes/range-large-v3/tests/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v3/tests/test.sh": "8fc106530f63500be0aace428a5dbf3e08567feb2adc40968029cb7d7a95afbc",
+    "labs/20-deltawire/eval/results/preflight-v3/index.json": "cab7fd3dffdf486d77b6ad7a144025d5c0b0bbd0708e823c9cf0c0261d6dfbe0",
+    "labs/20-deltawire/eval/results/preflight-v3/readiness.json": "26ae7263a63c96d6565131080c1b95dfe92b954b31c408003cdd9cf5d9bb9875",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-derived-assessment.json": "a802fc15372556197d3e274aff8e308c0208b396bdf11bb3656a6a360b384172",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-plan-contract-receipt.json": "dcf531cd8e86207a951aa9e43e80d45d7a9f8df5e3baa224feafbc2de3dd6fa0",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-semantic-result.json": "3e80260f930364f1b72255ed9898862e641536a8c6e84a75e9a4cbea3d1b46bb",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-treatment-fidelity.json": "394054b28824efc34eb26cad91c135b10933657cf6c8ec788c19b219b15cb5ad",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/config.json": "e8c95ea651adc99ac412322e2e08b735ef002fdab283a1e9c2a2f968698fd60c",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/job.log": "689ee87a1b8ced2b1c159103d7a329c91b889c1e971637a428ee0847ba03412b",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/lock.json": "4fdb9e97b806a06eb7a952310e25329e9dd30b768392ce4062903b83f8e4a048",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/agent/gemini-cli.trajectory.jsonl": "069cfca7ba46baf5757e809de50e53a656973f5b0a0fbe597a1cf763dce4cd27",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/agent/gemini-cli.txt": "54f08f9e9112fe5b46cb1f17fa6866496b58293f07d3b65a96ce82d0b6eb38f6",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/agent/trajectory.json": "6ddaaa2b6b582fd56f600a29bf78fb6bf74d1bf39d2bb668c0b00793fa2e0a68",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/artifacts/.deltawire/config.json": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/artifacts/.deltawire/plan-contract-receipt.json": "dcf531cd8e86207a951aa9e43e80d45d7a9f8df5e3baa224feafbc2de3dd6fa0",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/artifacts/.deltawire/plans/range-large.dw.json": "2021238a6573d8cf11477fb1d2656b9586b84415335edf69be4ff8fc74ef3a77",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/artifacts/.deltawire/schemas/range-large.schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/artifacts/manifest.json": "639ca0ff42307921ded25177e8d0343df7be9621011c7235f6593835f78a4188",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/artifacts/testdata/generated/output.ndjson": "9d205b57ab166d23c81927899f99b6ccece261c679c96979c58fbde975637776",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/config.json": "c6d8111f06134293971b16d7690b5f81b5a43c4d8356c04bd650fdb29a1f4a3c",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/lock.json": "66447eea0f175e275254a4f87ed377b6409328860335b917a2a390bd74e47a16",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/result.json": "e787ad6ecf1f83810124e0d9e0e1a63d829bbd6d587d2b5d9586735ecf89c068",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/trial.log": "689ee87a1b8ced2b1c159103d7a329c91b889c1e971637a428ee0847ba03412b",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/verifier/reward.json": "36a492513204040f04d7bdf4146d84ede71353d8cc7df112fc3ebff66327683a",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/verifier/test-stdout.txt": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/result.json": "27356b5f2bee385d3d9ee465dbef77a518c10d76e7348deb99578eb395a1d7a3",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/redaction.json": "afd1361d1db4b96aae9fbfdc972d80c67aa9d490567417bc47ed68bff29a62ca",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/report.md": "b3cae2d8e3239d4667eb36dcd1c658fc1def98c4031fa6a6d252af8a3df65b41",
+    "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/run-ledger.json": "3bcdc70b82afdfcf6b2200cba7042032d6f24abbf55dc7f148438d2e4234d99e",
+    "labs/20-deltawire/eval/scripts/fidelity.py": "f5bd8c7bf1ae7f3d99f681f1c2edf9ffb3a7bface84e28fc96a9f2ec2bffd350",
+    "labs/20-deltawire/eval/scripts/generate_preflight_v3_manifest.py": "a3b79335eac24626fddeb91b874151d42996ef9accfb40b72407c16c28f00515",
+    "labs/20-deltawire/eval/scripts/runner.py": "ba28350a7546a7284834aed702a5934585b50cc4534a519c09332a1e28e17825",
+    "labs/20-deltawire/eval/scripts/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/scripts/task_contracts.py": "e72514979104f539d431fec7f400d2067a45c0d0814bc054bc937ff113c1fee2",
+    "labs/20-deltawire/eval/scripts/test_fidelity.py": "90708e8f2c29b53fc2c5b6a3e2f3738ba179f96c2503e257445c3ff2fdf6bab3",
+    "labs/20-deltawire/eval/scripts/test_runner_gates.py": "9c070d68893782362548b281eeccb4bdb6110a75624dee63a18b1e48948d278b",
+    "labs/20-deltawire/eval/scripts/validate_preflight.py": "3220c8ecd8fb5139871fa7b93a64497efbea5c8ae57280a207184168f44ff159",
+    "labs/20-deltawire/eval/scripts/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/skills/deltawire/SKILL.md": "b1c5ed8e1abd0bf4e2031d887657bbfbc79204eed5c9d48a97fa1fbf50c5695f",
+    "labs/20-deltawire/eval/tasks/range-large/authoritative-schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/config.json": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+    "labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/schemas/range-large.schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/tasks/range-large/environment/Dockerfile": "0d4281bcb20ed360a4815c6170b1bd3fa65beaf51942d7203809b075ca4d3ead",
+    "labs/20-deltawire/eval/tasks/range-large/environment/task-contract.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/tasks/range-large/environment/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/tasks/range-large/instruction.md": "549f4047b1a4ba4f7a3f3dce9e60d1d70b141caa189aafdba942b15e6d022aee",
+    "labs/20-deltawire/eval/tasks/range-large/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/tasks/range-large/task.toml": "b7d4eff84bfbab08d514e1f39b58cdaaa1e925f4b28e09ec1e160041a45ef6b8",
+    "labs/20-deltawire/eval/tasks/range-large/tests/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/tasks/range-large/tests/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/tasks/range-large/tests/test.sh": "8fc106530f63500be0aace428a5dbf3e08567feb2adc40968029cb7d7a95afbc"
+  },
+  "schema_version": "git-evidence-lock.v1",
+  "source_commit": "07e341fd213b6deb17cce9115263cebe8ba2105f",
+  "source_tree": "cfad619db42d4426e464836d0e2be2de8384ddac"
+}
diff --git a/labs/20-deltawire/eval/manifests/preflight-v3-range-large.json b/labs/20-deltawire/eval/manifests/preflight-v3-range-large.json
new file mode 100644
index 000000000..e71511445
--- /dev/null
+++ b/labs/20-deltawire/eval/manifests/preflight-v3-range-large.json
@@ -0,0 +1,57 @@
+{
+  "agent": "gemini-cli",
+  "deltawire_binary_sha256": "453819efd1afe30b7e1815507de67b3150374955970a77231ce89a7a738c33c6",
+  "deltawire_version": "deltawire version dev",
+  "expected_reported_model": "gemini-3.1-pro-preview",
+  "failure_policy": "One v3 probe only; release one range-large/r1 pair only after raw-evidence rechecks; never run full.",
+  "harbor_command": [
+    "uvx",
+    "--from",
+    "harbor==0.20.0",
+    "harbor"
+  ],
+  "harbor_package": "harbor==0.20.0",
+  "harbor_version": "0.20.0",
+  "input_hashes": {
+    "labs/20-deltawire/eval/manifests/preflight-v2-evidence-lock.json": "dfaa4350681a36bcadef1e00d52bd90e16b2cf72638928ba2d0bc42b7af8f1b8",
+    "labs/20-deltawire/eval/matrices/72-run-matrix.json": "f9dcddc4b7336418c7e2b8e7a7e51ca1e42b5a137255ff0823a732618f3568d5",
+    "labs/20-deltawire/eval/probes/range-large-v3/authoritative-schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/probes/range-large-v3/environment/.deltawire/config.json": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+    "labs/20-deltawire/eval/probes/range-large-v3/environment/.deltawire/schemas/range-large.schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/probes/range-large-v3/environment/Dockerfile": "0d4281bcb20ed360a4815c6170b1bd3fa65beaf51942d7203809b075ca4d3ead",
+    "labs/20-deltawire/eval/probes/range-large-v3/environment/task-contract.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v3/environment/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/probes/range-large-v3/instruction.md": "2c1086bd55056b63c89bbb359fdb7a32168ce6c9598d02d6c084a52ad46e9af9",
+    "labs/20-deltawire/eval/probes/range-large-v3/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v3/task.toml": "854a1379edee3d090dfe85265753bcd045d192ccbd21e71a2003ac037436a49a",
+    "labs/20-deltawire/eval/probes/range-large-v3/tests/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/probes/range-large-v3/tests/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v3/tests/test.sh": "8fc106530f63500be0aace428a5dbf3e08567feb2adc40968029cb7d7a95afbc",
+    "labs/20-deltawire/eval/scripts/fidelity.py": "f5bd8c7bf1ae7f3d99f681f1c2edf9ffb3a7bface84e28fc96a9f2ec2bffd350",
+    "labs/20-deltawire/eval/scripts/runner.py": "ba28350a7546a7284834aed702a5934585b50cc4534a519c09332a1e28e17825",
+    "labs/20-deltawire/eval/scripts/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/scripts/task_contracts.py": "e72514979104f539d431fec7f400d2067a45c0d0814bc054bc937ff113c1fee2",
+    "labs/20-deltawire/eval/scripts/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/skills/deltawire/SKILL.md": "b1c5ed8e1abd0bf4e2031d887657bbfbc79204eed5c9d48a97fa1fbf50c5695f",
+    "labs/20-deltawire/eval/tasks/range-large/authoritative-schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/config.json": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+    "labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/schemas/range-large.schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/tasks/range-large/environment/Dockerfile": "0d4281bcb20ed360a4815c6170b1bd3fa65beaf51942d7203809b075ca4d3ead",
+    "labs/20-deltawire/eval/tasks/range-large/environment/task-contract.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/tasks/range-large/environment/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/tasks/range-large/instruction.md": "549f4047b1a4ba4f7a3f3dce9e60d1d70b141caa189aafdba942b15e6d022aee",
+    "labs/20-deltawire/eval/tasks/range-large/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/tasks/range-large/task.toml": "b7d4eff84bfbab08d514e1f39b58cdaaa1e925f4b28e09ec1e160041a45ef6b8",
+    "labs/20-deltawire/eval/tasks/range-large/tests/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/tasks/range-large/tests/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/tasks/range-large/tests/test.sh": "8fc106530f63500be0aace428a5dbf3e08567feb2adc40968029cb7d7a95afbc"
+  },
+  "model": "google/gemini-3.1-pro-preview",
+  "ready_for_72": false,
+  "repository_commit": "b82306a3",
+  "schema_version": "deltawire-preflight-v3-manifest.v1",
+  "scope": [
+    "probe/range-large-v3",
+    "range-large/r1"
+  ]
+}
diff --git a/labs/20-deltawire/eval/manifests/preflight-v4-range-large.json b/labs/20-deltawire/eval/manifests/preflight-v4-range-large.json
new file mode 100644
index 000000000..c2599c92f
--- /dev/null
+++ b/labs/20-deltawire/eval/manifests/preflight-v4-range-large.json
@@ -0,0 +1,84 @@
+{
+  "agent": "gemini-cli",
+  "deltawire_binary_sha256": "d50785f7479a5d143c548051e1972b6fb6edacba768824cd9c28561c51bd3b5a",
+  "expected_reported_model": "gemini-3.1-pro-preview",
+  "failure_policy": "One v4 probe; one automatically released range-large/r1 pair only after probe-release.v2; never rerun or run full.",
+  "harbor_command": [
+    "uvx",
+    "--from",
+    "harbor==0.20.0",
+    "harbor"
+  ],
+  "harbor_package": "harbor==0.20.0",
+  "harbor_version": "0.20.0",
+  "input_hashes": {
+    "labs/20-deltawire/eval/manifests/preflight-v2-evidence-lock.json": "dfaa4350681a36bcadef1e00d52bd90e16b2cf72638928ba2d0bc42b7af8f1b8",
+    "labs/20-deltawire/eval/manifests/preflight-v3-evidence-lock.json": "3318137e8f0da487297c3e95d07bac61fb7229d4ff751c334fe444e4444d6f61",
+    "labs/20-deltawire/eval/matrices/72-run-matrix.json": "f9dcddc4b7336418c7e2b8e7a7e51ca1e42b5a137255ff0823a732618f3568d5",
+    "labs/20-deltawire/eval/probes/range-large-v4/authoritative-schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/probes/range-large-v4/environment/.deltawire/config.json": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+    "labs/20-deltawire/eval/probes/range-large-v4/environment/.deltawire/schemas/range-large.schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/probes/range-large-v4/environment/Dockerfile": "10f480e7735bab6ca1f9389af058951550284516a98c363b6eb113d4f9ff655f",
+    "labs/20-deltawire/eval/probes/range-large-v4/environment/deltawire-environment-expectations.json": "5b2c2f4876ef27633b8889ce4bc9209a8233b0ca2c7dfefca24b61de5cabb252",
+    "labs/20-deltawire/eval/probes/range-large-v4/environment/environment_receipt.py": "bf1a730b409faa8ee8a85a9a83a2f076dc2c1daad72f25c9613e6f18d630a90f",
+    "labs/20-deltawire/eval/probes/range-large-v4/environment/task-contract.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v4/environment/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/probes/range-large-v4/instruction.md": "e71b24acdf77f476573e4373a7f6a814f94e35adf5dcd0e6f048350574359725",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/authoritative-schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.deltawire/config.json": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.deltawire/schemas/range-large.schema.json": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/Dockerfile": "10f480e7735bab6ca1f9389af058951550284516a98c363b6eb113d4f9ff655f",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/deltawire-environment-expectations.json": "5b2c2f4876ef27633b8889ce4bc9209a8233b0ca2c7dfefca24b61de5cabb252",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/environment_receipt.py": "bf1a730b409faa8ee8a85a9a83a2f076dc2c1daad72f25c9613e6f18d630a90f",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/task-contract.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/instruction.md": "549f4047b1a4ba4f7a3f3dce9e60d1d70b141caa189aafdba942b15e6d022aee",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/task.toml": "01f68e68f24bd143fc2cf9cce1c89871694160a56992d1c19ef03f6f02dc1fc9",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/test.sh": "0c5eb4de20a4c76fb073742a684f15ff3431b38bfa0a9ca27186e4b1f9e65884",
+    "labs/20-deltawire/eval/probes/range-large-v4/skill/deltawire/SKILL.md": "69f266e086759f3b8371a5708464067d10daf30c7d01a8c8bc06dd39261ec19e",
+    "labs/20-deltawire/eval/probes/range-large-v4/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v4/task.toml": "81f4033d57800432e053aa9166b887f98fff6a4ca339753a5575917ecbe42bd9",
+    "labs/20-deltawire/eval/probes/range-large-v4/tests/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/probes/range-large-v4/tests/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+    "labs/20-deltawire/eval/probes/range-large-v4/tests/test.sh": "0c5eb4de20a4c76fb073742a684f15ff3431b38bfa0a9ca27186e4b1f9e65884",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/v1-evidence-lock.json": "26f4013270eb366c3c190f8235ef2564985b593005d499338273b1edc2c051d1",
+    "labs/20-deltawire/eval/scripts/v4/build_and_stage.sh": "ad93bff7e05e1c6f40ffc8663ba688c77bced2a26c5bf11ccfda35ffc312af16",
+    "labs/20-deltawire/eval/scripts/v4/environment_receipt.py": "bf1a730b409faa8ee8a85a9a83a2f076dc2c1daad72f25c9613e6f18d630a90f",
+    "labs/20-deltawire/eval/scripts/v4/fidelity.py": "ae1963ea7021dc61eb38da757706ffe0dae204a96e757f452663080529f74077",
+    "labs/20-deltawire/eval/scripts/v4/generate_manifest.py": "f2223056e66ceb660b8954f9bc6cb7b32291819dd2ca0573608041bc1d5d8e2b",
+    "labs/20-deltawire/eval/scripts/v4/generate_v3_evidence_lock.py": "dfccde6c65d9357adf320bc8c28a7a5ad0341b3579b5505126d8e3b5ce3f7b43",
+    "labs/20-deltawire/eval/scripts/v4/runner.py": "789cd527a7e3659d77f5879347d5493784d30bb5d8ca02c9ed5cdcdb62312547",
+    "labs/20-deltawire/eval/scripts/v4/semantic_oracle.py": "94b30fac117f40e977851966b5228e6c372bb48cbf2e5c29c24f428e1f98e91d",
+    "labs/20-deltawire/eval/scripts/v4/shell_observation.py": "e5269e06350147715505a76b8e83235abecd5af34eba1c8837e19b22b0e4c9fe",
+    "labs/20-deltawire/eval/scripts/v4/test_environment_receipt.py": "59421625f731a5c5c9e555d970b87c1977eda86253cd18382b5aef4679c59d9e",
+    "labs/20-deltawire/eval/scripts/v4/test_fidelity.py": "531ca3f45bad8c3896829c430e2a01743bd24ceeb102170b7efa57525d545c19",
+    "labs/20-deltawire/eval/scripts/v4/test_runner.py": "ca72def4dc39be2bfb9b0cbe8e5eb5031b3c8711008abe7473e361eaf2901696",
+    "labs/20-deltawire/eval/scripts/v4/test_shell_observation.py": "5348c52e60b362408b433813466eb8cd8f03bde8f80033359eadb502bb71f3a0",
+    "labs/20-deltawire/eval/scripts/v4/validate.py": "571b7d95ac23595fd66268189691644872f0d78eaaf1f6814f27dde50cde9e9f",
+    "labs/20-deltawire/eval/scripts/v4/validate.sh": "c2ab7b0270a773f36fe877adb69829f0584902105fc2f12512ad6375533f5558",
+    "labs/20-deltawire/eval/scripts/v4/verify_evidence_locks.py": "8422b5f1bb156bffe6e399586262b474a6ac4bae48cfafb58b348bb2be4eafa0",
+    "labs/20-deltawire/eval/scripts/v4/verify_plan_contract.py": "8aafe73ab47213027a258f9abbebbca4803fa53f658abeb7b405d4a79202d4af",
+    "labs/20-deltawire/eval/tasks/range-large/task-spec.json": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5"
+  },
+  "matrix_positions": {
+    "order": [
+      "D1",
+      "B0"
+    ],
+    "zero_based": [
+      36,
+      37
+    ]
+  },
+  "model": "google/gemini-3.1-pro-preview",
+  "ready_for_72": false,
+  "repository_commit": "5a8ee77989efd85733e6d0efef75aa3fe2a9af1c",
+  "schema_version": "deltawire-preflight-v4-manifest.v1",
+  "scope": [
+    "probe/range-large-v4",
+    "range-large/r1"
+  ]
+}
diff --git a/labs/20-deltawire/eval/matrices/72-run-matrix.json b/labs/20-deltawire/eval/matrices/72-run-matrix.json
new file mode 100644
index 000000000..de95b5232
--- /dev/null
+++ b/labs/20-deltawire/eval/matrices/72-run-matrix.json
@@ -0,0 +1,434 @@
+[
+  {
+    "pair_id": "rows-medium/r1",
+    "task": "rows-medium",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "rows-medium/r1",
+    "task": "rows-medium",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "irregular-large/r2",
+    "task": "irregular-large",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "irregular-large/r2",
+    "task": "irregular-large",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "mixed-medium/r1",
+    "task": "mixed-medium",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "mixed-medium/r1",
+    "task": "mixed-medium",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "matrix-medium/r3",
+    "task": "matrix-medium",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "matrix-medium/r3",
+    "task": "matrix-medium",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "rows-large/r3",
+    "task": "rows-large",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "rows-large/r3",
+    "task": "rows-large",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "range-medium/r2",
+    "task": "range-medium",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "range-medium/r2",
+    "task": "range-medium",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "range-large/r3",
+    "task": "range-large",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "range-large/r3",
+    "task": "range-large",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "irregular-medium/r3",
+    "task": "irregular-medium",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "irregular-medium/r3",
+    "task": "irregular-medium",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "variants-large/r2",
+    "task": "variants-large",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "variants-large/r2",
+    "task": "variants-large",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "matrix-large/r3",
+    "task": "matrix-large",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "matrix-large/r3",
+    "task": "matrix-large",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "range-medium/r3",
+    "task": "range-medium",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "range-medium/r3",
+    "task": "range-medium",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "variants-large/r1",
+    "task": "variants-large",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "variants-large/r1",
+    "task": "variants-large",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "variants-medium/r1",
+    "task": "variants-medium",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "variants-medium/r1",
+    "task": "variants-medium",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "mixed-large/r3",
+    "task": "mixed-large",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "mixed-large/r3",
+    "task": "mixed-large",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "rows-large/r1",
+    "task": "rows-large",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "rows-large/r1",
+    "task": "rows-large",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "rows-large/r2",
+    "task": "rows-large",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "rows-large/r2",
+    "task": "rows-large",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "mixed-large/r2",
+    "task": "mixed-large",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "mixed-large/r2",
+    "task": "mixed-large",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "variants-medium/r3",
+    "task": "variants-medium",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "variants-medium/r3",
+    "task": "variants-medium",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "range-large/r1",
+    "task": "range-large",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "range-large/r1",
+    "task": "range-large",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "irregular-large/r3",
+    "task": "irregular-large",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "irregular-large/r3",
+    "task": "irregular-large",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "irregular-large/r1",
+    "task": "irregular-large",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "irregular-large/r1",
+    "task": "irregular-large",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "variants-large/r3",
+    "task": "variants-large",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "variants-large/r3",
+    "task": "variants-large",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "irregular-medium/r2",
+    "task": "irregular-medium",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "irregular-medium/r2",
+    "task": "irregular-medium",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "range-large/r2",
+    "task": "range-large",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "range-large/r2",
+    "task": "range-large",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "matrix-large/r1",
+    "task": "matrix-large",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "matrix-large/r1",
+    "task": "matrix-large",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "rows-medium/r3",
+    "task": "rows-medium",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "rows-medium/r3",
+    "task": "rows-medium",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "mixed-medium/r3",
+    "task": "mixed-medium",
+    "repetition": 3,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "mixed-medium/r3",
+    "task": "mixed-medium",
+    "repetition": 3,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "rows-medium/r2",
+    "task": "rows-medium",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "rows-medium/r2",
+    "task": "rows-medium",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "matrix-medium/r1",
+    "task": "matrix-medium",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "matrix-medium/r1",
+    "task": "matrix-medium",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "variants-medium/r2",
+    "task": "variants-medium",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "variants-medium/r2",
+    "task": "variants-medium",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "mixed-large/r1",
+    "task": "mixed-large",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "mixed-large/r1",
+    "task": "mixed-large",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "matrix-large/r2",
+    "task": "matrix-large",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "matrix-large/r2",
+    "task": "matrix-large",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "matrix-medium/r2",
+    "task": "matrix-medium",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "matrix-medium/r2",
+    "task": "matrix-medium",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "mixed-medium/r2",
+    "task": "mixed-medium",
+    "repetition": 2,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "mixed-medium/r2",
+    "task": "mixed-medium",
+    "repetition": 2,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "irregular-medium/r1",
+    "task": "irregular-medium",
+    "repetition": 1,
+    "arm": "B0"
+  },
+  {
+    "pair_id": "irregular-medium/r1",
+    "task": "irregular-medium",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "range-medium/r1",
+    "task": "range-medium",
+    "repetition": 1,
+    "arm": "D1"
+  },
+  {
+    "pair_id": "range-medium/r1",
+    "task": "range-medium",
+    "repetition": 1,
+    "arm": "B0"
+  }
+]
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/authoritative-schema.json b/labs/20-deltawire/eval/probes/range-large-v2/authoritative-schema.json
new file mode 100644
index 000000000..69b476237
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/authoritative-schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 500,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/assets/init/config.json b/labs/20-deltawire/eval/probes/range-large-v2/environment/.deltawire/config.json
similarity index 64%
rename from labs/20-deltawire/assets/init/config.json
rename to labs/20-deltawire/eval/probes/range-large-v2/environment/.deltawire/config.json
index 3fc5bd827..61a05415a 100644
--- a/labs/20-deltawire/assets/init/config.json
+++ b/labs/20-deltawire/eval/probes/range-large-v2/environment/.deltawire/config.json
@@ -1,12 +1,12 @@
 {
-  "version": "deltawire.config.v1",
-  "plans_dir": ".deltawire/plans",
-  "schemas_dir": ".deltawire/schemas",
-  "state_file": ".deltawire/state.json",
   "limits": {
+    "max_output_bytes": 104857600,
     "max_plan_bytes": 1048576,
-    "max_schema_bytes": 1048576,
     "max_records": 100000,
-    "max_output_bytes": 104857600
-  }
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
 }
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/environment/.deltawire/schemas/range-large.schema.json b/labs/20-deltawire/eval/probes/range-large-v2/environment/.deltawire/schemas/range-large.schema.json
new file mode 100644
index 000000000..69b476237
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/environment/.deltawire/schemas/range-large.schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 500,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/environment/Dockerfile b/labs/20-deltawire/eval/probes/range-large-v2/environment/Dockerfile
new file mode 100644
index 000000000..126972c7a
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/environment/Dockerfile
@@ -0,0 +1,8 @@
+FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/environment/task-contract.json b/labs/20-deltawire/eval/probes/range-large-v2/environment/task-contract.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/environment/task-contract.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/environment/verify_plan_contract.py b/labs/20-deltawire/eval/probes/range-large-v2/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/instruction.md b/labs/20-deltawire/eval/probes/range-large-v2/instruction.md
new file mode 100644
index 000000000..573bff295
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/instruction.md
@@ -0,0 +1,28 @@
+This is the versioned non-benchmark DeltaWire treatment-fidelity probe. Run `gemini skills list`, activate the `deltawire` skill using the structured tool, and follow it exactly. Read public `/task-contract.json`; prove path, version, and SHA-256 of `/usr/local/bin/deltawire`; prove the immutable schema hash. Create only `.deltawire/plans/range-large.dw.json`, then validate, run `verify_plan_contract.py`, render, and check. Do not edit the schema or generate expanded records manually. Complete contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
+```
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/task-spec.json b/labs/20-deltawire/eval/probes/range-large-v2/task-spec.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/task.toml b/labs/20-deltawire/eval/probes/range-large-v2/task.toml
new file mode 100644
index 000000000..f5be26a21
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/task.toml
@@ -0,0 +1,23 @@
+schema_version = "1.3"
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/range-large.dw.json", ".deltawire/schemas/range-large.schema.json", ".deltawire/plan-contract-receipt.json"]
+
+[task]
+name = "operatorstack/range-large-treatment-probe-v2"
+description = "Non-benchmark external-contract probe"
+authors = []
+keywords = []
+
+[metadata]
+benchmark_result = false
+
+[verifier]
+timeout_sec = 600.0
+
+[agent]
+timeout_sec = 600.0
+
+[environment]
+network_mode = "public"
+build_timeout_sec = 600.0
+os = "linux"
+mcp_servers = []
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/tests/semantic_oracle.py b/labs/20-deltawire/eval/probes/range-large-v2/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/tests/task-spec.json b/labs/20-deltawire/eval/probes/range-large-v2/tests/task-spec.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/tests/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v2/tests/test.sh b/labs/20-deltawire/eval/probes/range-large-v2/tests/test.sh
new file mode 100644
index 000000000..9b56450e6
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v2/tests/test.sh
@@ -0,0 +1,7 @@
+#!/bin/bash
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/authoritative-schema.json b/labs/20-deltawire/eval/probes/range-large-v3/authoritative-schema.json
new file mode 100644
index 000000000..69b476237
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/authoritative-schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 500,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/environment/.deltawire/config.json b/labs/20-deltawire/eval/probes/range-large-v3/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/environment/.deltawire/schemas/range-large.schema.json b/labs/20-deltawire/eval/probes/range-large-v3/environment/.deltawire/schemas/range-large.schema.json
new file mode 100644
index 000000000..69b476237
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/environment/.deltawire/schemas/range-large.schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 500,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/environment/Dockerfile b/labs/20-deltawire/eval/probes/range-large-v3/environment/Dockerfile
new file mode 100644
index 000000000..126972c7a
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/environment/Dockerfile
@@ -0,0 +1,8 @@
+FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/environment/task-contract.json b/labs/20-deltawire/eval/probes/range-large-v3/environment/task-contract.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/environment/task-contract.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/environment/verify_plan_contract.py b/labs/20-deltawire/eval/probes/range-large-v3/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/instruction.md b/labs/20-deltawire/eval/probes/range-large-v3/instruction.md
new file mode 100644
index 000000000..0be5c300e
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/instruction.md
@@ -0,0 +1,28 @@
+This is the versioned range-large-v3 non-benchmark DeltaWire treatment-fidelity recovery probe. Before activation, run `gemini skills list` and retain its structured result. Then activate the `deltawire` skill using the structured tool and follow it exactly. Read public `/task-contract.json`; prove path, version, and SHA-256 of `/usr/local/bin/deltawire`; prove the immutable public contract, config, and schema hashes. Create only `.deltawire/plans/range-large.dw.json`, then run validate, `verify_plan_contract.py`, render, and check. Do not edit the config or schema and do not generate expanded records manually. Complete contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
+```
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/task-spec.json b/labs/20-deltawire/eval/probes/range-large-v3/task-spec.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/task.toml b/labs/20-deltawire/eval/probes/range-large-v3/task.toml
new file mode 100644
index 000000000..437b492a5
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/task.toml
@@ -0,0 +1,24 @@
+schema_version = "1.3"
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/range-large.dw.json", ".deltawire/schemas/range-large.schema.json", ".deltawire/plan-contract-receipt.json"]
+
+[task]
+name = "operatorstack/range-large-treatment-probe-v3"
+description = "Versioned non-benchmark external-contract recovery probe"
+authors = []
+keywords = []
+
+[metadata]
+benchmark_result = false
+probe_version = "v3"
+
+[verifier]
+timeout_sec = 600.0
+
+[agent]
+timeout_sec = 600.0
+
+[environment]
+network_mode = "public"
+build_timeout_sec = 600.0
+os = "linux"
+mcp_servers = []
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/tests/semantic_oracle.py b/labs/20-deltawire/eval/probes/range-large-v3/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/tests/task-spec.json b/labs/20-deltawire/eval/probes/range-large-v3/tests/task-spec.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/tests/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v3/tests/test.sh b/labs/20-deltawire/eval/probes/range-large-v3/tests/test.sh
new file mode 100644
index 000000000..9b56450e6
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v3/tests/test.sh
@@ -0,0 +1,7 @@
+#!/bin/bash
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/authoritative-schema.json b/labs/20-deltawire/eval/probes/range-large-v4/authoritative-schema.json
new file mode 100644
index 000000000..69b476237
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/authoritative-schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 500,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/environment/.deltawire/config.json b/labs/20-deltawire/eval/probes/range-large-v4/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/environment/.deltawire/schemas/range-large.schema.json b/labs/20-deltawire/eval/probes/range-large-v4/environment/.deltawire/schemas/range-large.schema.json
new file mode 100644
index 000000000..69b476237
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/environment/.deltawire/schemas/range-large.schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 500,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/environment/Dockerfile b/labs/20-deltawire/eval/probes/range-large-v4/environment/Dockerfile
new file mode 100644
index 000000000..023396b14
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/environment/Dockerfile
@@ -0,0 +1,10 @@
+FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+COPY environment_receipt.py /usr/local/bin/environment_receipt.py
+COPY deltawire-environment-expectations.json /deltawire-environment-expectations.json
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py /usr/local/bin/environment_receipt.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json /deltawire-environment-expectations.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/environment/deltawire-environment-expectations.json b/labs/20-deltawire/eval/probes/range-large-v4/environment/deltawire-environment-expectations.json
new file mode 100644
index 000000000..a1d40d93d
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/environment/deltawire-environment-expectations.json
@@ -0,0 +1,14 @@
+{
+  "schema_version": "deltawire-environment-expectations.v1",
+  "binary": {
+    "path": "/usr/local/bin/deltawire",
+    "realpath": "/usr/local/bin/deltawire",
+    "sha256": "d50785f7479a5d143c548051e1972b6fb6edacba768824cd9c28561c51bd3b5a",
+    "version": "deltawire version dev"
+  },
+  "files": {
+    "config": {"path": "/.deltawire/config.json", "sha256": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f"},
+    "schema": {"path": "/.deltawire/schemas/range-large.schema.json", "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"},
+    "public_contract": {"path": "/task-contract.json", "sha256": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5"}
+  }
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/environment/environment_receipt.py b/labs/20-deltawire/eval/probes/range-large-v4/environment/environment_receipt.py
new file mode 100644
index 000000000..b973414ba
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/environment/environment_receipt.py
@@ -0,0 +1,36 @@
+#!/usr/bin/env python3
+"""Generate deterministic DeltaWire infrastructure evidence inside the task."""
+import argparse,hashlib,json,os,stat,subprocess
+from pathlib import Path
+
+def sha(path):
+ h=hashlib.sha256()
+ with Path(path).open("rb") as f:
+  for chunk in iter(lambda:f.read(1048576),b""): h.update(chunk)
+ return h.hexdigest()
+def main():
+ p=argparse.ArgumentParser(); p.add_argument("--expectations",required=True); p.add_argument("--receipt",required=True); a=p.parse_args()
+ expected=json.loads(Path(a.expectations).read_text()); binary=Path(expected["binary"]["path"])
+ exists=binary.exists(); lst=binary.lstat() if exists else None
+ is_regular=bool(lst and stat.S_ISREG(lst.st_mode)); is_symlink=binary.is_symlink() if exists else False
+ executable=exists and os.access(binary,os.X_OK); realpath=str(binary.resolve()) if exists else None
+ try: version=subprocess.run([str(binary),"version"],capture_output=True,text=True,check=False) if exists else None
+ except OSError: version=None
+ files={}
+ for name,item in expected["files"].items():
+  path=Path(item["path"]); actual=sha(path) if path.is_file() else None
+  files[name]={"path":item["path"],"exists":path.is_file(),"actual_sha256":actual,"expected_sha256":item["sha256"],"hash_match":actual==item["sha256"]}
+ binary_hash=sha(binary) if is_regular else None
+ checks={
+  "binary_exists":exists,"binary_regular":is_regular,"binary_not_symlink":not is_symlink,
+  "binary_executable":executable,"binary_path":str(binary)==expected["binary"]["path"],
+  "binary_realpath":realpath==expected["binary"]["realpath"],
+  "binary_hash":binary_hash==expected["binary"]["sha256"],
+  "version_exit_0":bool(version and version.returncode==0),
+  "version_exact":bool(version and version.stdout.splitlines() and version.stdout.splitlines()[0].strip()==expected["binary"]["version"]),
+  **{f"{name}_hash":item["hash_match"] for name,item in files.items()},
+ }
+ receipt={"schema_version":"deltawire-environment-receipt.v1","binary":{"path":str(binary),"realpath":realpath,"exists":exists,"is_regular":is_regular,"is_symlink":is_symlink,"executable":executable,"actual_sha256":binary_hash,"expected_sha256":expected["binary"]["sha256"],"version_stdout":version.stdout.strip() if version else None,"version_stderr":version.stderr.strip() if version else None,"version_exit_code":version.returncode if version else None,"expected_version":expected["binary"]["version"]},"files":files,"checks":checks,"status":"pass" if all(checks.values()) else "fail"}
+ target=Path(a.receipt); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(receipt,indent=2,sort_keys=True)+"\n")
+ print(json.dumps(receipt,sort_keys=True)); raise SystemExit(0 if receipt["status"]=="pass" else 1)
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/environment/task-contract.json b/labs/20-deltawire/eval/probes/range-large-v4/environment/task-contract.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/environment/task-contract.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/environment/verify_plan_contract.py b/labs/20-deltawire/eval/probes/range-large-v4/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/instruction.md b/labs/20-deltawire/eval/probes/range-large-v4/instruction.md
new file mode 100644
index 000000000..9f50b55d7
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/instruction.md
@@ -0,0 +1,28 @@
+This is the versioned range-large-v4 non-benchmark measurement-recovery probe. Before activation, run `gemini skills list` and retain its structured result. Then activate the `deltawire` skill using the structured tool and follow it exactly. Read public `/task-contract.json`. Create only `.deltawire/plans/range-large.dw.json`; do not create or edit config, schema, state, environment receipts, or expanded output manually. Run validate, `verify_plan_contract.py`, render, and check. Infrastructure path, version, executability, and hashes are verified independently outside model reasoning. Complete contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
+```
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/authoritative-schema.json b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/authoritative-schema.json
new file mode 100644
index 000000000..69b476237
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/authoritative-schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 500,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.deltawire/config.json b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.deltawire/schemas/range-large.schema.json b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.deltawire/schemas/range-large.schema.json
new file mode 100644
index 000000000..69b476237
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.deltawire/schemas/range-large.schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 500,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.generated/.gitignore b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.generated/.gitignore
new file mode 100644
index 000000000..d6b7ef32c
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/.generated/.gitignore
@@ -0,0 +1,2 @@
+*
+!.gitignore
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/Dockerfile b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/Dockerfile
new file mode 100644
index 000000000..023396b14
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/Dockerfile
@@ -0,0 +1,10 @@
+FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+COPY environment_receipt.py /usr/local/bin/environment_receipt.py
+COPY deltawire-environment-expectations.json /deltawire-environment-expectations.json
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py /usr/local/bin/environment_receipt.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json /deltawire-environment-expectations.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/deltawire-environment-expectations.json b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/deltawire-environment-expectations.json
new file mode 100644
index 000000000..a1d40d93d
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/deltawire-environment-expectations.json
@@ -0,0 +1,14 @@
+{
+  "schema_version": "deltawire-environment-expectations.v1",
+  "binary": {
+    "path": "/usr/local/bin/deltawire",
+    "realpath": "/usr/local/bin/deltawire",
+    "sha256": "d50785f7479a5d143c548051e1972b6fb6edacba768824cd9c28561c51bd3b5a",
+    "version": "deltawire version dev"
+  },
+  "files": {
+    "config": {"path": "/.deltawire/config.json", "sha256": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f"},
+    "schema": {"path": "/.deltawire/schemas/range-large.schema.json", "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"},
+    "public_contract": {"path": "/task-contract.json", "sha256": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5"}
+  }
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/environment_receipt.py b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/environment_receipt.py
new file mode 100644
index 000000000..b973414ba
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/environment_receipt.py
@@ -0,0 +1,36 @@
+#!/usr/bin/env python3
+"""Generate deterministic DeltaWire infrastructure evidence inside the task."""
+import argparse,hashlib,json,os,stat,subprocess
+from pathlib import Path
+
+def sha(path):
+ h=hashlib.sha256()
+ with Path(path).open("rb") as f:
+  for chunk in iter(lambda:f.read(1048576),b""): h.update(chunk)
+ return h.hexdigest()
+def main():
+ p=argparse.ArgumentParser(); p.add_argument("--expectations",required=True); p.add_argument("--receipt",required=True); a=p.parse_args()
+ expected=json.loads(Path(a.expectations).read_text()); binary=Path(expected["binary"]["path"])
+ exists=binary.exists(); lst=binary.lstat() if exists else None
+ is_regular=bool(lst and stat.S_ISREG(lst.st_mode)); is_symlink=binary.is_symlink() if exists else False
+ executable=exists and os.access(binary,os.X_OK); realpath=str(binary.resolve()) if exists else None
+ try: version=subprocess.run([str(binary),"version"],capture_output=True,text=True,check=False) if exists else None
+ except OSError: version=None
+ files={}
+ for name,item in expected["files"].items():
+  path=Path(item["path"]); actual=sha(path) if path.is_file() else None
+  files[name]={"path":item["path"],"exists":path.is_file(),"actual_sha256":actual,"expected_sha256":item["sha256"],"hash_match":actual==item["sha256"]}
+ binary_hash=sha(binary) if is_regular else None
+ checks={
+  "binary_exists":exists,"binary_regular":is_regular,"binary_not_symlink":not is_symlink,
+  "binary_executable":executable,"binary_path":str(binary)==expected["binary"]["path"],
+  "binary_realpath":realpath==expected["binary"]["realpath"],
+  "binary_hash":binary_hash==expected["binary"]["sha256"],
+  "version_exit_0":bool(version and version.returncode==0),
+  "version_exact":bool(version and version.stdout.splitlines() and version.stdout.splitlines()[0].strip()==expected["binary"]["version"]),
+  **{f"{name}_hash":item["hash_match"] for name,item in files.items()},
+ }
+ receipt={"schema_version":"deltawire-environment-receipt.v1","binary":{"path":str(binary),"realpath":realpath,"exists":exists,"is_regular":is_regular,"is_symlink":is_symlink,"executable":executable,"actual_sha256":binary_hash,"expected_sha256":expected["binary"]["sha256"],"version_stdout":version.stdout.strip() if version else None,"version_stderr":version.stderr.strip() if version else None,"version_exit_code":version.returncode if version else None,"expected_version":expected["binary"]["version"]},"files":files,"checks":checks,"status":"pass" if all(checks.values()) else "fail"}
+ target=Path(a.receipt); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(receipt,indent=2,sort_keys=True)+"\n")
+ print(json.dumps(receipt,sort_keys=True)); raise SystemExit(0 if receipt["status"]=="pass" else 1)
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/task-contract.json b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/task-contract.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/task-contract.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/verify_plan_contract.py b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/instruction.md b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/instruction.md
new file mode 100644
index 000000000..9181f4398
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/instruction.md
@@ -0,0 +1,28 @@
+Generate the exact `range-large` dataset defined by public `/task-contract.json`. Write exactly 500 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
+```
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/task-spec.json b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/task-spec.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/task.toml b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/task.toml
new file mode 100644
index 000000000..21f7117db
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/task.toml
@@ -0,0 +1,24 @@
+schema_version = "1.3"
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/range-large.dw.json", ".deltawire/schemas/range-large.schema.json", ".deltawire/state.json", ".deltawire/plan-contract-receipt.json", ".deltawire/environment-receipt.json"]
+
+[task]
+name = "operatorstack/range-large-v4-pair"
+description = "Frozen range-large/r1 operational positive-control task"
+authors = []
+keywords = []
+
+[metadata]
+benchmark_result = true
+probe_version = "v4-pair"
+
+[verifier]
+timeout_sec = 600.0
+
+[agent]
+timeout_sec = 600.0
+
+[environment]
+network_mode = "public"
+build_timeout_sec = 600.0
+os = "linux"
+mcp_servers = []
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/semantic_oracle.py b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/task-spec.json b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/task-spec.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/test.sh b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/test.sh
new file mode 100644
index 000000000..1dc0f71db
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/pair-task/tests/test.sh
@@ -0,0 +1,9 @@
+#!/bin/bash
+mkdir -p /logs/verifier
+python3 /usr/local/bin/environment_receipt.py --expectations /deltawire-environment-expectations.json --receipt /.deltawire/environment-receipt.json || true
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/skill/deltawire/SKILL.md b/labs/20-deltawire/eval/probes/range-large-v4/skill/deltawire/SKILL.md
new file mode 100644
index 000000000..22776d5af
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/skill/deltawire/SKILL.md
@@ -0,0 +1,28 @@
+---
+name: deltawire
+description: Use for deterministic generation of large repetitive JSON or NDJSON datasets from compact Cartesian products, ranges, repeated rows, variants, or regular-plus-exception patterns. Discover with `gemini skills list` before activation. Abstain for small or highly irregular outputs.
+---
+
+# DeltaWire
+
+Before activation, use a structured shell call to run `gemini skills list` and
+confirm `deltawire` is listed. Activate this skill only when the requested data
+has a compact deterministic description. Explicitly abstain when it is too
+irregular for a compact deterministic plan.
+
+Read `/task-contract.json`. Treat its output path, record count, ordering,
+schema path, schema hash, and assertions as authoritative. The preinstalled
+config and schema are immutable. Create only
+`.deltawire/plans/range-large.dw.json`; never create or edit config, schema,
+state, environment receipts, verifier files, or expanded output manually.
+
+Run the lifecycle in order:
+
+1. `deltawire validate --repo . .deltawire/plans/range-large.dw.json`
+2. `verify_plan_contract.py --contract /task-contract.json --repo . --plan .deltawire/plans/range-large.dw.json --receipt .deltawire/plan-contract-receipt.json`
+3. `deltawire render --repo . .deltawire/plans/range-large.dw.json`
+4. `deltawire check --repo . .deltawire/plans/range-large.dw.json`
+
+The independent environment receipt, retained-artifact contract check, and
+semantic oracle are the final authorities. Do not infer their results or
+replace them with prose.
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/task-spec.json b/labs/20-deltawire/eval/probes/range-large-v4/task-spec.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/task.toml b/labs/20-deltawire/eval/probes/range-large-v4/task.toml
new file mode 100644
index 000000000..d12fe1f1b
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/task.toml
@@ -0,0 +1,24 @@
+schema_version = "1.3"
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/range-large.dw.json", ".deltawire/schemas/range-large.schema.json", ".deltawire/state.json", ".deltawire/plan-contract-receipt.json", ".deltawire/environment-receipt.json"]
+
+[task]
+name = "operatorstack/range-large-treatment-probe-v4"
+description = "Versioned measurement-boundary recovery probe"
+authors = []
+keywords = []
+
+[metadata]
+benchmark_result = false
+probe_version = "v4"
+
+[verifier]
+timeout_sec = 600.0
+
+[agent]
+timeout_sec = 600.0
+
+[environment]
+network_mode = "public"
+build_timeout_sec = 600.0
+os = "linux"
+mcp_servers = []
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/tests/semantic_oracle.py b/labs/20-deltawire/eval/probes/range-large-v4/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/tests/task-spec.json b/labs/20-deltawire/eval/probes/range-large-v4/tests/task-spec.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/tests/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/probes/range-large-v4/tests/test.sh b/labs/20-deltawire/eval/probes/range-large-v4/tests/test.sh
new file mode 100644
index 000000000..1dc0f71db
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large-v4/tests/test.sh
@@ -0,0 +1,9 @@
+#!/bin/bash
+mkdir -p /logs/verifier
+python3 /usr/local/bin/environment_receipt.py --expectations /deltawire-environment-expectations.json --receipt /.deltawire/environment-receipt.json || true
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/probes/range-large/environment/Dockerfile b/labs/20-deltawire/eval/probes/range-large/environment/Dockerfile
new file mode 100644
index 000000000..b2504dafa
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large/environment/Dockerfile
@@ -0,0 +1,5 @@
+FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+RUN chmod 0755 /usr/local/bin/deltawire
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/probes/range-large/instruction.md b/labs/20-deltawire/eval/probes/range-large/instruction.md
new file mode 100644
index 000000000..0fd52b5a8
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large/instruction.md
@@ -0,0 +1 @@
+This is a treatment-delivery probe, not a benchmark run. First use the shell to run `gemini skills list` and prove that `deltawire` is discovered. Then activate the `deltawire` skill using the structured skill activation tool. Follow the activated skill exactly. Prove `command -v deltawire`, `deltawire version`, and the SHA-256 of `/usr/local/bin/deltawire`. Use DeltaWire to generate the exact dataset in `tests/task-spec.json` at `testdata/generated/output.ndjson`, retaining the schema at `.deltawire/schemas/range-large.schema.json` and plan at `.deltawire/plans/range-large.dw.json`. Run `deltawire validate`, `deltawire render`, and `deltawire check`; do not generate expanded data manually.
diff --git a/labs/20-deltawire/eval/probes/range-large/task.toml b/labs/20-deltawire/eval/probes/range-large/task.toml
new file mode 100644
index 000000000..ffa146fbf
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large/task.toml
@@ -0,0 +1,23 @@
+schema_version = "1.3"
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/plans/range-large.dw.json", ".deltawire/schemas/range-large.schema.json"]
+
+[task]
+name = "operatorstack/range-large-treatment-probe"
+description = "Non-benchmark DeltaWire treatment delivery probe"
+authors = []
+keywords = []
+
+[metadata]
+benchmark_result = false
+
+[verifier]
+timeout_sec = 600.0
+
+[agent]
+timeout_sec = 600.0
+
+[environment]
+network_mode = "public"
+build_timeout_sec = 600.0
+os = "linux"
+mcp_servers = []
diff --git a/labs/20-deltawire/eval/probes/range-large/tests/semantic_oracle.py b/labs/20-deltawire/eval/probes/range-large/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/probes/range-large/tests/task-spec.json b/labs/20-deltawire/eval/probes/range-large/tests/task-spec.json
new file mode 100644
index 000000000..b57098325
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large/tests/task-spec.json
@@ -0,0 +1,20 @@
+{
+  "schema_version": "task-spec.v1",
+  "task": "range-large",
+  "family": "range",
+  "size": "large",
+  "output": {
+    "path": "testdata/generated/output.ndjson",
+    "format": "ndjson"
+  },
+  "record_count": 500,
+  "ordering": "canonical_generation_order",
+  "deltawire_applicability": "expected",
+  "generation": {
+    "start": 1,
+    "field": "index",
+    "field_order": [
+      "index"
+    ]
+  }
+}
diff --git a/labs/20-deltawire/eval/probes/range-large/tests/test.sh b/labs/20-deltawire/eval/probes/range-large/tests/test.sh
new file mode 100755
index 000000000..a4935826b
--- /dev/null
+++ b/labs/20-deltawire/eval/probes/range-large/tests/test.sh
@@ -0,0 +1,5 @@
+#!/usr/bin/env bash
+OUTPUT_FILE="testdata/generated/output.ndjson"
+mkdir -p /logs/verifier
+if [ ! -f "$OUTPUT_FILE" ]; then echo '{"exact_match": 0}' > /logs/verifier/reward.json; exit 0; fi
+python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/results/README.md b/labs/20-deltawire/eval/results/README.md
index a5ca95f0c..6501f0832 100644
--- a/labs/20-deltawire/eval/results/README.md
+++ b/labs/20-deltawire/eval/results/README.md
@@ -1 +1,11 @@
 # Results
+
+- `preflight-v1/matrix-medium-r1`: completed overhead-control pair; excluded
+  from confirmatory estimates.
+- `preflight-v1/treatment-probe-range-large`: historical failed semantic probe.
+- `preflight-v2/treatment-probe-range-large-v2`: external-contract probe;
+  semantic pass, delivery/end-to-end fail, no pair launched.
+- `preflight-v3/treatment-probe-range-large-v3`: recovery probe; semantic and
+  retained-artifact checks passed, but the frozen parser rejected the proven
+  `which deltawire` path form; no release or pair was launched.
+- `archive/`: provenance-only historical and failed artifacts.
diff --git a/labs/20-deltawire/eval/results/archive/failed-harbor-v0/README.md b/labs/20-deltawire/eval/results/archive/failed-harbor-v0/README.md
new file mode 100644
index 000000000..2a4ad1dfd
--- /dev/null
+++ b/labs/20-deltawire/eval/results/archive/failed-harbor-v0/README.md
@@ -0,0 +1,5 @@
+# Archived failed Harbor preflight jobs
+
+These six tracked jobs predate the verified matrix-medium pair and failed with
+infrastructure or agent execution errors. They are retained only for provenance
+and are excluded from ledgers, analysis, manifests, and readiness calculations.
diff --git a/labs/20-deltawire/eval/results/archive/legacy-preflight-v0/6-run-canary.json b/labs/20-deltawire/eval/results/archive/legacy-preflight-v0/6-run-canary.json
new file mode 100644
index 000000000..e6ece27a3
--- /dev/null
+++ b/labs/20-deltawire/eval/results/archive/legacy-preflight-v0/6-run-canary.json
@@ -0,0 +1,38 @@
+[
+  {
+    "pair_id": "mixed-medium/r1",
+    "task": "mixed-medium",
+    "arm": "B0",
+    "repetition": 1
+  },
+  {
+    "pair_id": "mixed-medium/r1",
+    "task": "mixed-medium",
+    "arm": "D1",
+    "repetition": 1
+  },
+  {
+    "pair_id": "matrix-medium/r1",
+    "task": "matrix-medium",
+    "arm": "D1",
+    "repetition": 1
+  },
+  {
+    "pair_id": "matrix-medium/r1",
+    "task": "matrix-medium",
+    "arm": "B0",
+    "repetition": 1
+  },
+  {
+    "pair_id": "irregular-medium/r1",
+    "task": "irregular-medium",
+    "arm": "B0",
+    "repetition": 1
+  },
+  {
+    "pair_id": "irregular-medium/r1",
+    "task": "irregular-medium",
+    "arm": "D1",
+    "repetition": 1
+  }
+]
\ No newline at end of file
diff --git a/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/classification.json b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/classification.json
new file mode 100644
index 000000000..92c66d66b
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/classification.json
@@ -0,0 +1,12 @@
+{
+  "schema_version": "preflight-classification.v1",
+  "pair_id": "matrix-medium/r1",
+  "task_role": "small regular / overhead control",
+  "semantic_result": "both_pass",
+  "treatment_result": "skill_configured_deltawire_not_used",
+  "causal_conclusion": "no DeltaWire runtime comparison occurred",
+  "confirmatory_estimate_eligible": false,
+  "b0_input_plus_output_tokens": 24397,
+  "d1_input_plus_output_tokens": 39027,
+  "d1_relative_increase": 0.5997
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/extracted_runs.json b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/extracted_runs.json
new file mode 100644
index 000000000..dee6f6a76
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/extracted_runs.json
@@ -0,0 +1,110 @@
+[
+  {
+    "pair_id": "matrix-medium/r1",
+    "task": "matrix-medium",
+    "arm": "D1",
+    "repetition": 1,
+    "trial_id": "19aee6b3-f588-4917-a273-22db7956a16b",
+    "actual_model": "gemini-3.1-pro-preview",
+    "actual_agent": "gemini-cli",
+    "configured_model": "google/gemini-3.1-pro-preview",
+    "verifier_reward": 1,
+    "semantic_oracle_result": {
+      "exact_match": 1
+    },
+    "oracle_command": "python3 /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/eval/scripts/semantic_oracle.py matrix-medium /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/D1/2026-07-22__07-59-23/matrix-medium__ADNi43q/artifacts/testdata/generated/output.ndjson",
+    "oracle_implementation_sha": "2f45dbc017771f3d44b6f628c443f373d70833e9efd2c270e8415456eb375900",
+    "input_tokens": 37475,
+    "output_tokens": 1552,
+    "cache_tokens": 24230,
+    "token_json_pointers": {
+      "input_tokens": "/agent_result/n_input_tokens",
+      "output_tokens": "/agent_result/n_output_tokens",
+      "cache_tokens": "/agent_result/n_cache_tokens"
+    },
+    "steps": null,
+    "tool_calls": null,
+    "wall_time": 62.113306,
+    "infrastructure_failure": false,
+    "result_path": "/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/D1/2026-07-22__07-59-23/matrix-medium__ADNi43q/result.json",
+    "result_sha256": "00e402ebf6344d9ca10c295025aee5a4f79b2b950bfde6049836e2b61af93e66",
+    "transcript_path": "/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/D1/2026-07-22__07-59-23/matrix-medium__ADNi43q/agent/gemini-cli.trajectory.jsonl",
+    "transcript_sha256": "3021cf293057747286b4305d35c438d922fcb8dc81d3d514c4f22d6833679479",
+    "output_artifact_path": "/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/D1/2026-07-22__07-59-23/matrix-medium__ADNi43q/artifacts/testdata/generated/output.ndjson",
+    "output_artifact_sha256": "a8e95911587aeef0d6bab401201eeb3f0c37c66e23d7a0b67a996beddda9ed82",
+    "output_byte_size": 323,
+    "record_count": 6,
+    "deltawire_compliance": "not_used",
+    "deltawire_mechanism": {
+      "skill_configured": true,
+      "validate_invoked": false,
+      "validate_exit_status": null,
+      "render_invoked": false,
+      "render_exit_status": null,
+      "check_invoked": false,
+      "check_exit_status": null,
+      "plan_path": null,
+      "plan_sha256": null,
+      "plan_bytes": null,
+      "unexpected_use": false
+    },
+    "executed_by_runner": true,
+    "redacted_command": "uvx harbor run -a gemini-cli -m google/gemini-3.1-pro-preview -p /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/eval/tasks/matrix-medium -k 1 --jobs-dir /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/D1 --ae 'GEMINI_API_KEY=***' --skill /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/eval/skills/deltawire",
+    "job_dir": "/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/D1/2026-07-22__07-59-23",
+    "trial_dir": "/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/D1/2026-07-22__07-59-23/matrix-medium__ADNi43q"
+  },
+  {
+    "pair_id": "matrix-medium/r1",
+    "task": "matrix-medium",
+    "arm": "B0",
+    "repetition": 1,
+    "trial_id": "86ab2589-688e-4026-9588-30d108e42021",
+    "actual_model": "gemini-3.1-pro-preview",
+    "actual_agent": "gemini-cli",
+    "configured_model": "google/gemini-3.1-pro-preview",
+    "verifier_reward": 1,
+    "semantic_oracle_result": {
+      "exact_match": 1
+    },
+    "oracle_command": "python3 /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/eval/scripts/semantic_oracle.py matrix-medium /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/B0/2026-07-22__08-00-52/matrix-medium__aG22NEn/artifacts/testdata/generated/output.ndjson",
+    "oracle_implementation_sha": "2f45dbc017771f3d44b6f628c443f373d70833e9efd2c270e8415456eb375900",
+    "input_tokens": 23582,
+    "output_tokens": 815,
+    "cache_tokens": 16154,
+    "token_json_pointers": {
+      "input_tokens": "/agent_result/n_input_tokens",
+      "output_tokens": "/agent_result/n_output_tokens",
+      "cache_tokens": "/agent_result/n_cache_tokens"
+    },
+    "steps": null,
+    "tool_calls": null,
+    "wall_time": 76.142906,
+    "infrastructure_failure": false,
+    "result_path": "/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/B0/2026-07-22__08-00-52/matrix-medium__aG22NEn/result.json",
+    "result_sha256": "c3e124ba0b6bfa92fb343f74cdf15ac226d8ffbcf5ec65e7c319730c8ee6fedb",
+    "transcript_path": "/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/B0/2026-07-22__08-00-52/matrix-medium__aG22NEn/agent/gemini-cli.trajectory.jsonl",
+    "transcript_sha256": "a496b44a54b3117d130c1aa574e61bbb7e06dc413e62957b13f8537bd9f13e06",
+    "output_artifact_path": "/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/B0/2026-07-22__08-00-52/matrix-medium__aG22NEn/artifacts/testdata/generated/output.ndjson",
+    "output_artifact_sha256": "cf7c41040a5a2df6efb7294ebeefbb425912942ae36ac4807709425106eff766",
+    "output_byte_size": 293,
+    "record_count": 6,
+    "deltawire_compliance": "not_used",
+    "deltawire_mechanism": {
+      "skill_configured": false,
+      "validate_invoked": false,
+      "validate_exit_status": null,
+      "render_invoked": false,
+      "render_exit_status": null,
+      "check_invoked": false,
+      "check_exit_status": null,
+      "plan_path": null,
+      "plan_sha256": null,
+      "plan_bytes": null,
+      "unexpected_use": false
+    },
+    "executed_by_runner": true,
+    "redacted_command": "uvx harbor run -a gemini-cli -m google/gemini-3.1-pro-preview -p /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/eval/tasks/matrix-medium -k 1 --jobs-dir /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/B0 --ae 'GEMINI_API_KEY=***'",
+    "job_dir": "/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/B0/2026-07-22__08-00-52",
+    "trial_dir": "/Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/B0/2026-07-22__08-00-52/matrix-medium__aG22NEn"
+  }
+]
diff --git a/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/paired_analysis.json b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/paired_analysis.json
new file mode 100644
index 000000000..4fbff3d63
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/paired_analysis.json
@@ -0,0 +1,18 @@
+{
+  "aggregate_reduction": -0.599663893101611,
+  "median_paired_reduction": -0.5996638931016108,
+  "geometric_mean_paired_reduction": -0.599663893101611,
+  "task_clustered_95_interval": [
+    -0.5996638931016108,
+    -0.5996638931016108
+  ],
+  "pair_wins_losses_ties": {
+    "wins": 0,
+    "losses": 1,
+    "ties": 0
+  },
+  "per_task_family_results": {},
+  "all_assigned_run_result": 2,
+  "both_pass_pair_result": 1,
+  "d1_mechanism_compliance_rate": 0.0
+}
\ No newline at end of file
diff --git a/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/readiness.json b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/readiness.json
new file mode 100644
index 000000000..ba763dd46
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/readiness.json
@@ -0,0 +1,9 @@
+{
+  "READY_FOR_72": false,
+  "reasons": [
+    "matrix-medium is an overhead control and D1 did not invoke DeltaWire",
+    "treatment fidelity has not yet been established"
+  ],
+  "pair_id": "matrix-medium/r1",
+  "confirmatory_estimate_eligible": false
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/report.md b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/report.md
new file mode 100644
index 000000000..9aa8dac53
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/report.md
@@ -0,0 +1,14 @@
+# Matrix-medium r1 Overhead Control
+
+One real Harbor B0/D1 pair completed. Both arms passed the independent semantic
+oracle. D1 had the DeltaWire skill configured but did not activate it or invoke
+the DeltaWire runtime. D1 reported 39,027 input-plus-output tokens versus
+24,397 for B0, approximately 60% more.
+
+This is a small regular overhead control, not treatment-fidelity evidence and
+not part of any confirmatory estimate. No DeltaWire runtime comparison occurred.
+`READY_FOR_72` remains false.
+
+All paths in `run_ledger.json` are repository-relative. The retained raw B0 and
+D1 Harbor results, trajectories, generated outputs, and verifier results live
+under this directory's `raw/` tree.
diff --git a/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/run_ledger.json b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/run_ledger.json
new file mode 100644
index 000000000..1ed80a03e
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/run_ledger.json
@@ -0,0 +1,44 @@
+{
+  "runs": [
+    {
+      "pair_id": "matrix-medium/r1",
+      "task": "matrix-medium",
+      "arm": "D1",
+      "repetition": 1,
+      "status": "completed",
+      "jobs_dir": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/D1",
+      "redacted_command": "uvx harbor run -a gemini-cli -m google/gemini-3.1-pro-preview -p /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/eval/tasks/matrix-medium -k 1 --jobs-dir /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/D1 --ae 'GEMINI_API_KEY=***' --skill /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/eval/skills/deltawire",
+      "executed_by_runner": true,
+      "harbor_returncode": 0,
+      "trial_id": "19aee6b3-f588-4917-a273-22db7956a16b",
+      "job_dir": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/D1/2026-07-22__07-59-23",
+      "trial_dir": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/D1/2026-07-22__07-59-23/matrix-medium__ADNi43q",
+      "result_path": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/D1/2026-07-22__07-59-23/matrix-medium__ADNi43q/result.json",
+      "result_sha256": "00e402ebf6344d9ca10c295025aee5a4f79b2b950bfde6049836e2b61af93e66",
+      "transcript_path": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/D1/2026-07-22__07-59-23/matrix-medium__ADNi43q/agent/gemini-cli.trajectory.jsonl",
+      "transcript_sha256": "3021cf293057747286b4305d35c438d922fcb8dc81d3d514c4f22d6833679479",
+      "output_artifact_path": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/D1/2026-07-22__07-59-23/matrix-medium__ADNi43q/artifacts/testdata/generated/output.ndjson",
+      "output_artifact_sha256": "a8e95911587aeef0d6bab401201eeb3f0c37c66e23d7a0b67a996beddda9ed82"
+    },
+    {
+      "pair_id": "matrix-medium/r1",
+      "task": "matrix-medium",
+      "arm": "B0",
+      "repetition": 1,
+      "status": "completed",
+      "jobs_dir": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/B0",
+      "redacted_command": "uvx harbor run -a gemini-cli -m google/gemini-3.1-pro-preview -p /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/eval/tasks/matrix-medium -k 1 --jobs-dir /Users/apple/Documents/GitHub/intelligence-flow/.product-loop/worktrees/revector/labs/20-deltawire/harbor/jobs/B0 --ae 'GEMINI_API_KEY=***'",
+      "executed_by_runner": true,
+      "harbor_returncode": 0,
+      "trial_id": "86ab2589-688e-4026-9588-30d108e42021",
+      "job_dir": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/B0/2026-07-22__08-00-52",
+      "trial_dir": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/B0/2026-07-22__08-00-52/matrix-medium__aG22NEn",
+      "result_path": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/B0/2026-07-22__08-00-52/matrix-medium__aG22NEn/result.json",
+      "result_sha256": "c3e124ba0b6bfa92fb343f74cdf15ac226d8ffbcf5ec65e7c319730c8ee6fedb",
+      "transcript_path": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/B0/2026-07-22__08-00-52/matrix-medium__aG22NEn/agent/gemini-cli.trajectory.jsonl",
+      "transcript_sha256": "a496b44a54b3117d130c1aa574e61bbb7e06dc413e62957b13f8537bd9f13e06",
+      "output_artifact_path": "labs/20-deltawire/eval/results/preflight-v1/matrix-medium-r1/raw/B0/2026-07-22__08-00-52/matrix-medium__aG22NEn/artifacts/testdata/generated/output.ndjson",
+      "output_artifact_sha256": "cf7c41040a5a2df6efb7294ebeefbb425912942ae36ac4807709425106eff766"
+    }
+  ]
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v1/readiness.json b/labs/20-deltawire/eval/results/preflight-v1/readiness.json
new file mode 100644
index 000000000..9a814e383
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/readiness.json
@@ -0,0 +1,10 @@
+{
+  "READY_FOR_72": false,
+  "reasons": [
+    "confirmatory 72-run experiment has not been authorized or executed",
+    "range-large treatment-fidelity probe did not pass all delivery and semantic gates",
+    "range-large/r1 positive-control pair incomplete"
+  ],
+  "probe_passed": false,
+  "positive_control_pair_passed": false
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/mechanism-delivery-receipt.legacy.json b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/mechanism-delivery-receipt.legacy.json
new file mode 100644
index 000000000..02a97e384
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/mechanism-delivery-receipt.legacy.json
@@ -0,0 +1,23 @@
+{
+  "schema_version": "treatment-fidelity-receipt.v1",
+  "classification": "used_successfully",
+  "discovered": true,
+  "skills_list_discovered": true,
+  "activated": true,
+  "binary_available": true,
+  "binary_sha256": "453819efd1afe30b7e1815507de67b3150374955970a77231ce89a7a738c33c6",
+  "binary_hash_match": true,
+  "validate_exit_0": true,
+  "render_exit_0": true,
+  "check_exit_0": true,
+  "plan_path": "artifacts/.deltawire/plans/range-large.dw.json",
+  "plan_sha256": "cb5daa4ce22a58ddd1f5354e79040b7fe95f2c5811b61152f5bbcb9e3ea8a2fc",
+  "plan_bytes": 620,
+  "semantic_exact_match": 0,
+  "trajectory_sha256": "4d1090eed6998ce68e4f8d5a1027e48945d27d3c396ed81ad20164767bd89266",
+  "explicit_abstention": false,
+  "result_sha256": "c8a360c19e6623eedc19ac9772043f8b6a5c5011b4e0c8f94a188f5b099c173a",
+  "output_sha256": "49f78bbc6fe581fd322b07cb0a6c682ebb929c364695183400042925880077a5",
+  "input_tokens": 1600539,
+  "output_tokens": 20036
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/redaction.json b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/redaction.json
new file mode 100644
index 000000000..acc47f8de
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/redaction.json
@@ -0,0 +1,16 @@
+{
+  "schema_version": "artifact-redaction.v1",
+  "reason": "Remove GEMINI_API_KEY values before committing Harbor evidence",
+  "files": [
+    {
+      "path": "raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/agent/trajectory.json",
+      "original_sha256": "df10cfa28d936924c8e064aba37bb85e82ca694890c310cb20ff1c5f9f8bdc54",
+      "redacted_sha256": "a97989ead95ebbb2521160968d5143c843590d98c62df24a72a7e9a5401444b6"
+    },
+    {
+      "path": "raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/agent/gemini-cli.trajectory.jsonl",
+      "original_sha256": "4d1090eed6998ce68e4f8d5a1027e48945d27d3c396ed81ad20164767bd89266",
+      "redacted_sha256": "bf41bd8561bd0395c7301bed85c5dcb3fdcf05b8074300cbbde1c833d304c428"
+    }
+  ]
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/report.md b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/report.md
new file mode 100644
index 000000000..542736449
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/report.md
@@ -0,0 +1,33 @@
+# Range-large Treatment-Fidelity Probe
+
+Canonical derived status: `delivery_status=pass`, `semantic_status=fail`, and
+`end_to_end_status=fail`. The original mechanism receipt is retained
+byte-for-byte as `mechanism-delivery-receipt.legacy.json`.
+
+This was one non-benchmark probe on the `tbench-c4d` Harbor VM. It was not
+rerun and is excluded from benchmark estimates.
+
+## Delivery evidence
+
+- Gemini discovered DeltaWire with `gemini skills list`.
+- The structured `activate_skill` call succeeded.
+- `/usr/local/bin/deltawire` was available and matched frozen SHA-256
+  `453819efd1afe30b7e1815507de67b3150374955970a77231ce89a7a738c33c6`.
+- `deltawire validate`, `render`, and `check` succeeded.
+- The retained 620-byte plan has SHA-256
+  `cb5daa4ce22a58ddd1f5354e79040b7fe95f2c5811b61152f5bbcb9e3ea8a2fc`.
+
+## Failed semantic gate
+
+The independent task oracle returned `exact_match=0`. The generated output had
+999 records with fields `id` and `value`; the frozen range-large contract
+requires 500 records with the sole field `index`, ordered from 1 through 500.
+The agent initially attempted to read a relative `tests/task-spec.json` path,
+while Harbor exposes verifier inputs under `/tests/`.
+
+Treatment delivery was demonstrated, but end-to-end probe fidelity was not.
+The required stop gate prevented the `range-large/r1` B0/D1 pair and the
+72-run experiment from running. `READY_FOR_72` remains false.
+
+Committed Harbor artifacts are credential-redacted; `redaction.json` retains
+the original and committed hashes for each changed file.
diff --git a/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/run-ledger.json b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/run-ledger.json
new file mode 100644
index 000000000..e72b1611b
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/run-ledger.json
@@ -0,0 +1,22 @@
+{
+  "schema_version": "run-ledger.v2",
+  "mode": "probe",
+  "runs": [
+    {
+      "pair_id": "probe/range-large",
+      "task": "range-large-treatment-probe",
+      "contract_task": "range-large",
+      "arm": "D1",
+      "repetition": 1,
+      "status": "completed",
+      "redacted_command": "uvx --from harbor==0.20.0 harbor run -a gemini-cli -m google/gemini-3.1-pro-preview -p /home/apple/deltawire-preflight/labs/20-deltawire/eval/probes/range-large -k 1 --jobs-dir /home/apple/deltawire-preflight/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1 --ae 'GEMINI_API_KEY=***' --skill /home/apple/deltawire-preflight/labs/20-deltawire/eval/skills/deltawire",
+      "trial_dir": "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB",
+      "result_path": "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/result.json",
+      "trajectory_path": "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/agent/gemini-cli.trajectory.jsonl",
+      "output_path": "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/artifacts/testdata/generated/output.ndjson",
+      "result_sha256": "c8a360c19e6623eedc19ac9772043f8b6a5c5011b4e0c8f94a188f5b099c173a",
+      "trajectory_sha256": "4d1090eed6998ce68e4f8d5a1027e48945d27d3c396ed81ad20164767bd89266",
+      "output_sha256": "49f78bbc6fe581fd322b07cb0a6c682ebb929c364695183400042925880077a5"
+    }
+  ]
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/treatment-fidelity-receipt.json b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/treatment-fidelity-receipt.json
new file mode 100644
index 000000000..6edfc1f69
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/treatment-fidelity-receipt.json
@@ -0,0 +1,11 @@
+{
+  "schema_version": "treatment-fidelity-receipt.v2",
+  "probe_version": "v1",
+  "delivery_status": "pass",
+  "semantic_status": "fail",
+  "end_to_end_status": "fail",
+  "delivery_evidence": {"skills_list_discovered": true, "activated": true, "binary_hash_match": true, "validate_exit_0": true, "render_exit_0": true, "check_exit_0": true, "plan_retained": true},
+  "semantic_exact_match": 0,
+  "legacy_mechanism_receipt": "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/mechanism-delivery-receipt.legacy.json",
+  "conclusion": "Mechanism delivery passed, but semantic and end-to-end fidelity failed."
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/v1-evidence-lock.json b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/v1-evidence-lock.json
new file mode 100644
index 000000000..ddbdd9c7d
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/v1-evidence-lock.json
@@ -0,0 +1,24 @@
+{
+  "immutable_files": {
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/mechanism-delivery-receipt.legacy.json": "00678e77671a966716ea76908860be779a1cbe3b7b09b3c4860f1b22f85c0628",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/config.json": "f9a074180dc34d06d0b07f50b61fd2e34429e0e3a632cde670fe4a1c99b58b9e",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/lock.json": "0a3e4ae39f8eebbef83862b2f00830b6273fcf448c5e05f306a3f8de7cc354f7",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/agent/gemini-cli.trajectory.jsonl": "bf41bd8561bd0395c7301bed85c5dcb3fdcf05b8074300cbbde1c833d304c428",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/agent/gemini-cli.txt": "742bc81b22fc6fe94ba6c6cb0136f8418eb8f881a939c9731848f0bc5ca97c32",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/agent/trajectory.json": "a97989ead95ebbb2521160968d5143c843590d98c62df24a72a7e9a5401444b6",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/artifacts/.deltawire/plans/range-large.dw.json": "cb5daa4ce22a58ddd1f5354e79040b7fe95f2c5811b61152f5bbcb9e3ea8a2fc",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/artifacts/.deltawire/schemas/range-large.schema.json": "db7d8fef5b4305a4a9a271dca62d6bb93a3a217c2ca31cf3796e116e5919a49e",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/artifacts/manifest.json": "fd43ed1d279f063de84809c3ef6e0b631efa82b47ff4d698f58d2ab42f803eb3",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/artifacts/testdata/generated/output.ndjson": "49f78bbc6fe581fd322b07cb0a6c682ebb929c364695183400042925880077a5",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/config.json": "e0874eae2b8df81c69b202a600c006826f800bc1f306643fbddc0266665220eb",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/lock.json": "d9d88f5cee36939443abacce164bdd8f72ab10cd3b19a11eea23e58e441874ef",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/result.json": "c8a360c19e6623eedc19ac9772043f8b6a5c5011b4e0c8f94a188f5b099c173a",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/verifier/reward.json": "05060fafdc7a05c47e0ae4fc8b0ecf217e7f758c76889b6eb52b1735a57a57ea",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/range-large__XBapLkB/verifier/test-stdout.txt": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/raw/D1/2026-07-22__07-44-54/result.json": "98367c584351a61ec0ccc72afd4aea658fbed74aedfbb8a0d01d5dd5eb03353b",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/redaction.json": "5d50edcb1b62c8fd8de05385aae8cb912151f67405f77f9e2fa1fd4fac70fbc0",
+    "labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large/run-ledger.json": "7d2db6e1b915a802cf046cadd36b5a94c8a81a8f08d1cd1381f6d18825d77e95"
+  },
+  "probe_version": "v1",
+  "schema_version": "evidence-lock.v1"
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v2/index.json b/labs/20-deltawire/eval/results/preflight-v2/index.json
new file mode 100644
index 000000000..e9bac9ec5
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v2/index.json
@@ -0,0 +1,15 @@
+{
+  "schema_version": "preflight-provenance-index.v1",
+  "entries": [
+    {
+      "id": "range-large-v2/D1",
+      "role": "delivery_probe",
+      "benchmark_result": false,
+      "path": "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2",
+      "delivery_status": "fail",
+      "semantic_status": "pass",
+      "end_to_end_status": "fail",
+      "pair_launched": false
+    }
+  ]
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v2/readiness.json b/labs/20-deltawire/eval/results/preflight-v2/readiness.json
new file mode 100644
index 000000000..669cbe559
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v2/readiness.json
@@ -0,0 +1,12 @@
+{
+  "READY_FOR_72": false,
+  "reasons": [
+    "six-run canary remains incomplete",
+    "final confirmatory manifest is not frozen",
+    "explicit 72-run authorization is absent",
+    "range-large treatment-fidelity probe did not pass all delivery and semantic gates",
+    "range-large/r1 positive-control pair incomplete"
+  ],
+  "probe_passed": false,
+  "positive_control_pair_passed": false
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/D1-plan-contract-receipt.json b/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/D1-plan-contract-receipt.json
new file mode 100644
index 000000000..38a73dc10
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/D1-plan-contract-receipt.json
@@ -0,0 +1,9 @@
+{
+  "checks": {},
+  "contract_sha256": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+  "error": "error: DW_CONFIG_INVALID: could not read config: open /home/apple/deltawire-preflight/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/range-large-v2__qCX5xEj/artifacts/.deltawire/config.json: no such file or directory",
+  "inspect_exit_code": 2,
+  "plan_sha256": "2d0b13459d191e62a02876b3404ae657037a4c0f853d3eb795140aed48e219dc",
+  "schema_version": "plan-contract-receipt.v1",
+  "status": "fail"
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/D1-treatment-fidelity.json b/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/D1-treatment-fidelity.json
new file mode 100644
index 000000000..9cc99a8e8
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/D1-treatment-fidelity.json
@@ -0,0 +1,36 @@
+{
+  "activated": true,
+  "binary_available": true,
+  "binary_hash_match": true,
+  "binary_sha256": "453819efd1afe30b7e1815507de67b3150374955970a77231ce89a7a738c33c6",
+  "check_exit_0": true,
+  "classification": "attempted_failed",
+  "contract_check_exit_0": true,
+  "delivery_status": "fail",
+  "discovered": true,
+  "end_to_end_status": "fail",
+  "explicit_abstention": false,
+  "failure_reasons": [
+    "No successful structured `gemini skills list` evidence was present.",
+    "Independent retained-artifact inspection failed because `.deltawire/config.json` was not retained."
+  ],
+  "independent_contract_check": "fail",
+  "independent_contract_receipt": "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/D1-plan-contract-receipt.json",
+  "input_tokens": 1396558,
+  "output_sha256": "9d205b57ab166d23c81927899f99b6ccece261c679c96979c58fbde975637776",
+  "output_tokens": 15308,
+  "plan_bytes": 383,
+  "plan_path": "artifacts/.deltawire/plans/range-large.dw.json",
+  "plan_sha256": "2d0b13459d191e62a02876b3404ae657037a4c0f853d3eb795140aed48e219dc",
+  "public_contract_accessed": true,
+  "render_exit_0": true,
+  "result_sha256": "4afb5cfaa8bf224a70ce82e4b152f2ff9b360ff8ac12079bd3b535fff71b149e",
+  "schema_hash_match": true,
+  "schema_sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+  "schema_version": "treatment-fidelity-receipt.v2",
+  "semantic_exact_match": 1,
+  "semantic_status": "pass",
+  "skills_list_discovered": false,
+  "trajectory_sha256": "f075b4ca6cf3352e5704d992f1ebcd09f50d3a42853675b0ab2ec82dc7977403",
+  "validate_exit_0": true
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/redaction.json b/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/redaction.json
new file mode 100644
index 000000000..1a3e33516
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/redaction.json
@@ -0,0 +1,7 @@
+{
+  "schema_version": "evidence-redaction.v1",
+  "credential_patterns_checked": ["Google API key", "GEMINI_API_KEY assignment"],
+  "unredacted_credentials_found": 0,
+  "run_ledger_assignment": "GEMINI_API_KEY=***",
+  "raw_evidence_committable": true
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/report.md b/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/report.md
new file mode 100644
index 000000000..c0f8e86e1
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/report.md
@@ -0,0 +1,20 @@
+# Range-large v2 External-Contract Probe
+
+This was exactly one D1-only, non-benchmark probe on `tbench-c4d`. Harbor's
+independent semantic verifier passed (`exact_match=1`): the output contains the
+canonical 500 integer `index` records. The in-container contract checker also
+passed and the authoritative schema hash was preserved.
+
+End-to-end fidelity nevertheless failed for two independently observable
+reasons. The transcript contains a successful structured `activate_skill` call
+but no successful `gemini skills list` discovery call. In addition, the
+runner's retained-artifact recheck could not initialize DeltaWire because the
+artifact list omitted `.deltawire/config.json`; its machine receipt is retained
+as `D1-plan-contract-receipt.json`.
+
+Status: `delivery_status=fail`, `semantic_status=pass`,
+`end_to_end_status=fail`. No B0 run and no `range-large/r1` pair were launched.
+Reported token fields are 1,396,558 input and 15,308 output tokens; they are
+probe diagnostics, not a savings claim.
+
+`READY_FOR_72` remains false.
diff --git a/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/run-ledger.json b/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/run-ledger.json
new file mode 100644
index 000000000..70c02544d
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/run-ledger.json
@@ -0,0 +1,22 @@
+{
+  "schema_version": "run-ledger.v2",
+  "mode": "probe-v2",
+  "runs": [
+    {
+      "pair_id": "probe/range-large-v2",
+      "task": "range-large-treatment-probe-v2",
+      "contract_task": "range-large",
+      "arm": "D1",
+      "repetition": 1,
+      "status": "completed",
+      "redacted_command": "uvx --from harbor==0.20.0 harbor run -a gemini-cli -m google/gemini-3.1-pro-preview -p /home/apple/deltawire-preflight/labs/20-deltawire/eval/probes/range-large-v2 -k 1 --jobs-dir /home/apple/deltawire-preflight/labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1 --ae 'GEMINI_API_KEY=***' --skill /home/apple/deltawire-preflight/labs/20-deltawire/eval/skills/deltawire",
+      "trial_dir": "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/range-large-v2__qCX5xEj",
+      "result_path": "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/range-large-v2__qCX5xEj/result.json",
+      "trajectory_path": "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/range-large-v2__qCX5xEj/agent/gemini-cli.trajectory.jsonl",
+      "output_path": "labs/20-deltawire/eval/results/preflight-v2/treatment-probe-range-large-v2/raw/D1/2026-07-22__11-05-50/range-large-v2__qCX5xEj/artifacts/testdata/generated/output.ndjson",
+      "result_sha256": "4afb5cfaa8bf224a70ce82e4b152f2ff9b360ff8ac12079bd3b535fff71b149e",
+      "trajectory_sha256": "f075b4ca6cf3352e5704d992f1ebcd09f50d3a42853675b0ab2ec82dc7977403",
+      "output_sha256": "9d205b57ab166d23c81927899f99b6ccece261c679c96979c58fbde975637776"
+    }
+  ]
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v3/index.json b/labs/20-deltawire/eval/results/preflight-v3/index.json
new file mode 100644
index 000000000..1ed7a9e95
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v3/index.json
@@ -0,0 +1,16 @@
+{
+  "schema_version": "preflight-provenance-index.v1",
+  "entries": [
+    {
+      "id": "range-large-v3/D1",
+      "role": "delivery_recovery_probe",
+      "benchmark_result": false,
+      "path": "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3",
+      "delivery_status": "fail",
+      "semantic_status": "pass",
+      "end_to_end_status": "fail",
+      "probe_release_created": false,
+      "pair_launched": false
+    }
+  ]
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v3/readiness.json b/labs/20-deltawire/eval/results/preflight-v3/readiness.json
new file mode 100644
index 000000000..4857f7ddc
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v3/readiness.json
@@ -0,0 +1,13 @@
+{
+  "READY_FOR_72": false,
+  "reasons": [
+    "six-run canary remains incomplete",
+    "final confirmatory manifest is not frozen",
+    "explicit 72-run authorization is absent",
+    "range-large-v3 did not produce a passing end-to-end treatment receipt",
+    "range-large/r1 was correctly blocked by the missing probe release"
+  ],
+  "probe_passed": false,
+  "probe_release_created": false,
+  "positive_control_pair_passed": false
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-derived-assessment.json b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-derived-assessment.json
new file mode 100644
index 000000000..a4d91d952
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-derived-assessment.json
@@ -0,0 +1,16 @@
+{
+  "schema_version": "treatment-fidelity-derived-assessment.v1",
+  "source_treatment_receipt": "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-treatment-fidelity.json",
+  "source_treatment_receipt_sha256": "394054b28824efc34eb26cad91c135b10933657cf6c8ec788c19b219b15cb5ad",
+  "source_trajectory_sha256": "069cfca7ba46baf5757e809de50e53a656973f5b0a0fbe597a1cf763dce4cd27",
+  "delivery_status": "fail",
+  "semantic_status": "pass",
+  "end_to_end_status": "fail",
+  "failure_reasons": [
+    "The frozen parser set binary_available=false because it recognizes only the literal command-v form.",
+    "The structured trajectory instead used which, returned /usr/local/bin/deltawire, and separately proved the expected version and SHA-256."
+  ],
+  "probe_release_created": false,
+  "pair_released": false,
+  "ready_for_72": false
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-plan-contract-receipt.json b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-plan-contract-receipt.json
new file mode 100644
index 000000000..efa410620
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-plan-contract-receipt.json
@@ -0,0 +1,39 @@
+{
+  "checks": {
+    "exact_count_assertion": true,
+    "output_format": true,
+    "output_path": true,
+    "projected_records": true,
+    "schema_exists": true,
+    "schema_path": true,
+    "schema_sha256": true
+  },
+  "contract_sha256": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+  "inspect": {
+    "assertion_summary": [
+      "exact count 500"
+    ],
+    "cmd_check": "deltawire check .deltawire/plans/range-large.dw.json",
+    "cmd_render": "deltawire render .deltawire/plans/range-large.dw.json",
+    "cmd_validate": "deltawire validate .deltawire/plans/range-large.dw.json",
+    "cold_amp_ratio": "6892/822",
+    "output_format": "ndjson",
+    "output_path": "testdata/generated/output.ndjson",
+    "plan_id": "",
+    "plan_only_amp_ratio": "6892/536",
+    "plan_path": ".deltawire/plans/range-large.dw.json",
+    "plan_source_bytes": 536,
+    "primitives_used": [
+      "matrix",
+      "ranges"
+    ],
+    "projected_output_bytes": 6892,
+    "projected_records": 500,
+    "schema_path": ".deltawire/schemas/range-large.schema.json",
+    "schema_source_bytes": 286
+  },
+  "inspect_exit_code": 0,
+  "plan_sha256": "2021238a6573d8cf11477fb1d2656b9586b84415335edf69be4ff8fc74ef3a77",
+  "schema_version": "plan-contract-receipt.v1",
+  "status": "pass"
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-semantic-result.json b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-semantic-result.json
new file mode 100644
index 000000000..c3d9fb823
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-semantic-result.json
@@ -0,0 +1,9 @@
+{
+  "exact_match": 1,
+  "oracle_result": {
+    "exact_match": 1
+  },
+  "schema_version": "semantic-result.v1",
+  "status": "pass",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-treatment-fidelity.json b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-treatment-fidelity.json
new file mode 100644
index 000000000..0d52cbeb6
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-treatment-fidelity.json
@@ -0,0 +1,56 @@
+{
+  "activated": true,
+  "artifact_set_complete": true,
+  "binary_available": false,
+  "binary_hash_match": true,
+  "binary_sha256": "453819efd1afe30b7e1815507de67b3150374955970a77231ce89a7a738c33c6",
+  "binary_version": "deltawire version dev",
+  "binary_version_proven": true,
+  "cache_tokens": 1323380,
+  "check_exit_0": true,
+  "classification": "attempted_failed",
+  "config_hash_match": true,
+  "config_hash_proven": true,
+  "config_retained": true,
+  "config_sha256": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+  "contract_check_exit_0": true,
+  "delivery_status": "fail",
+  "deltawire_shell_use": true,
+  "discovered": true,
+  "end_to_end_status": "fail",
+  "explicit_abstention": false,
+  "finished_at": "2026-07-22T12:05:47.064979Z",
+  "harbor_exit_0": true,
+  "in_container_contract_check": "pass",
+  "in_container_contract_receipt": "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/artifacts/.deltawire/plan-contract-receipt.json",
+  "in_container_contract_receipt_sha256": "dcf531cd8e86207a951aa9e43e80d45d7a9f8df5e3baa224feafbc2de3dd6fa0",
+  "independent_contract_check": "pass",
+  "independent_contract_receipt": "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/D1-plan-contract-receipt.json",
+  "input_tokens": 1581949,
+  "model_match": true,
+  "output_sha256": "9d205b57ab166d23c81927899f99b6ccece261c679c96979c58fbde975637776",
+  "output_tokens": 16017,
+  "plan_bytes": 536,
+  "plan_path": "artifacts/.deltawire/plans/range-large.dw.json",
+  "plan_sha256": "2021238a6573d8cf11477fb1d2656b9586b84415335edf69be4ff8fc74ef3a77",
+  "projected_record_count": 500,
+  "public_contract_accessed": true,
+  "public_contract_hash_match": true,
+  "public_contract_sha256": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+  "render_exit_0": true,
+  "reported_model": "gemini-3.1-pro-preview",
+  "result_sha256": "e787ad6ecf1f83810124e0d9e0e1a63d829bbd6d587d2b5d9586735ecf89c068",
+  "retained_schema_integrity": true,
+  "schema_hash_match": true,
+  "schema_sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+  "schema_version": "treatment-fidelity-receipt.v3",
+  "semantic_exact_match": 1,
+  "semantic_result_sha256": "3e80260f930364f1b72255ed9898862e641536a8c6e84a75e9a4cbea3d1b46bb",
+  "semantic_status": "pass",
+  "skills_list_discovered": true,
+  "started_at": "2026-07-22T12:00:56.309288Z",
+  "trajectory_sha256": "069cfca7ba46baf5757e809de50e53a656973f5b0a0fbe597a1cf763dce4cd27",
+  "trial_id": "7f366aab-8f8e-4edd-9752-db045e2660f9",
+  "validate_exit_0": true,
+  "wall_time_seconds": 290.755691
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/redaction.json b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/redaction.json
new file mode 100644
index 000000000..1a2e0e8e5
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/redaction.json
@@ -0,0 +1,14 @@
+{
+  "schema_version": "evidence-redaction.v1",
+  "credential_patterns_checked": [
+    "exact sourced Gemini credential bytes",
+    "Google API key",
+    "GEMINI_API_KEY assignment",
+    "GitHub token",
+    "private key block"
+  ],
+  "unredacted_credentials_found": 0,
+  "credential_transport": "private temporary env file removed after Harbor exited",
+  "run_ledger_assignment": "<private-env-file>",
+  "raw_evidence_committable": true
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/report.md b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/report.md
new file mode 100644
index 000000000..726cdb3c4
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/report.md
@@ -0,0 +1,28 @@
+# Range-large v3 Recovery Probe
+
+This directory preserves exactly one D1-only, non-benchmark v3 probe executed
+on `tbench-c4d`. Harbor 0.20.0 and Gemini 3.1 Pro Preview completed without an
+exception in 290.756 seconds. The independent semantic oracle passed
+(`exact_match=1`) for the canonical 500 integer `index` records.
+
+The intended delivery mechanism was present in the structured trajectory:
+`gemini skills list` listed DeltaWire before `activate_skill`; activation
+succeeded; the public contract, config, and schema hashes matched; validate,
+the in-container contract check, render, and check succeeded; the config,
+schema, plan, contract receipt, and output were retained; and the independent
+retained-artifact contract recheck passed with a projected count of 500.
+
+The frozen fidelity parser nevertheless recorded `binary_available=false`.
+The structured command used `which deltawire` and returned
+`/usr/local/bin/deltawire`, while the parser recognizes only the literal
+`command -v deltawire` form. Version and binary SHA-256 proofs separately
+passed. Because the frozen runner is fail-closed, this parser false negative
+made `delivery_status=fail` and `end_to_end_status=fail`; no
+`probe-release.v1` was created.
+
+The authorized `range-large/r1` pair was therefore not launched. Reported
+probe diagnostics are 1,581,949 input tokens, 16,017 output tokens, 1,323,380
+cache tokens, and 290.756 seconds wall time. They are not a benchmark comparison
+or a savings claim.
+
+`READY_FOR_72` remains false. No v4 probe or 72-run execution is authorized.
diff --git a/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/run-ledger.json b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/run-ledger.json
new file mode 100644
index 000000000..37a930384
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/run-ledger.json
@@ -0,0 +1,32 @@
+{
+  "mode": "probe-v3",
+  "ready_for_72": false,
+  "runs": [
+    {
+      "arm": "D1",
+      "cache_tokens": 1323380,
+      "contract_task": "range-large",
+      "finished_at": "2026-07-22T12:05:47.064979Z",
+      "harbor_exit_code": 0,
+      "input_tokens": 1581949,
+      "output_path": "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/artifacts/testdata/generated/output.ndjson",
+      "output_sha256": "9d205b57ab166d23c81927899f99b6ccece261c679c96979c58fbde975637776",
+      "output_tokens": 16017,
+      "pair_id": "probe/range-large-v3",
+      "redacted_command": "uvx --from harbor==0.20.0 harbor run -a gemini-cli -m google/gemini-3.1-pro-preview -p /home/apple/deltawire-preflight/labs/20-deltawire/eval/probes/range-large-v3 -k 1 --jobs-dir /home/apple/deltawire-preflight/labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1 --n-concurrent 1 --skill /home/apple/deltawire-preflight/labs/20-deltawire/eval/skills/deltawire --env-file '<private-env-file>'",
+      "repetition": 1,
+      "reported_model": "gemini-3.1-pro-preview",
+      "result_path": "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/result.json",
+      "result_sha256": "e787ad6ecf1f83810124e0d9e0e1a63d829bbd6d587d2b5d9586735ecf89c068",
+      "started_at": "2026-07-22T12:00:56.309288Z",
+      "status": "completed",
+      "task": "range-large-treatment-probe-v3",
+      "trajectory_path": "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht/agent/gemini-cli.trajectory.jsonl",
+      "trajectory_sha256": "069cfca7ba46baf5757e809de50e53a656973f5b0a0fbe597a1cf763dce4cd27",
+      "trial_dir": "labs/20-deltawire/eval/results/preflight-v3/treatment-probe-range-large-v3/raw/D1/2026-07-22__12-00-56/range-large-v3__wx2HCht",
+      "trial_id": "7f366aab-8f8e-4edd-9752-db045e2660f9",
+      "wall_time_seconds": 290.755691
+    }
+  ],
+  "schema_version": "run-ledger.v3"
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v4/index.json b/labs/20-deltawire/eval/results/preflight-v4/index.json
new file mode 100644
index 000000000..c5c61bf1f
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/index.json
@@ -0,0 +1,17 @@
+{
+  "schema_version": "preflight-provenance-index.v1",
+  "entries": [
+    {
+      "id": "range-large-v4/D1",
+      "role": "measurement_recovery_probe",
+      "benchmark_result": false,
+      "path": "labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4",
+      "environment_status": "fail",
+      "treatment_use_status": "pass",
+      "semantic_status": "pass",
+      "end_to_end_status": "fail",
+      "probe_release_created": false,
+      "pair_launched": false
+    }
+  ]
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v4/integration-projection.json b/labs/20-deltawire/eval/results/preflight-v4/integration-projection.json
new file mode 100644
index 000000000..bcb9df44b
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/integration-projection.json
@@ -0,0 +1,24 @@
+{
+  "schema_version": "deltawire-integration-projection.v1",
+  "source_branch": "eval/deltawire-preflight-v4",
+  "evidence_commit": "c200dc0eb3c4dff0e4732cc4b2daa1cd535703e3",
+  "evidence_pull_request": 99,
+  "excluded_path_segments": [
+    "/raw/"
+  ],
+  "excluded_artifact_classes": [
+    "raw trajectories",
+    "Harbor logs",
+    "expanded generated output",
+    "other raw run evidence"
+  ],
+  "included_artifact_classes": [
+    "implementation",
+    "contracts and oracles",
+    "manifests and locks",
+    "compact receipts",
+    "readiness records",
+    "reports"
+  ],
+  "READY_FOR_72": false
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v4/readiness.json b/labs/20-deltawire/eval/results/preflight-v4/readiness.json
new file mode 100644
index 000000000..6a750a825
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/readiness.json
@@ -0,0 +1,13 @@
+{
+  "READY_FOR_72": false,
+  "reasons": [
+    "six-run canary remains incomplete",
+    "final confirmatory manifest is not frozen",
+    "explicit 72-run authorization is absent",
+    "range-large-v4 did not retain its verifier-generated environment receipt",
+    "range-large/r1 was correctly blocked by the missing probe-release.v2"
+  ],
+  "probe_passed": false,
+  "probe_release_created": false,
+  "positive_control_pair_passed": false
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-end-to-end.json b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-end-to-end.json
new file mode 100644
index 000000000..f4a12064e
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-end-to-end.json
@@ -0,0 +1,19 @@
+{
+  "arm": "D1",
+  "environment_status": "fail",
+  "harbor_exit_0": true,
+  "model_match": true,
+  "ready_for_72": false,
+  "schema_version": "deltawire-end-to-end-receipt.v4",
+  "semantic_status": "pass",
+  "source_hashes": {
+    "environment": "c088201283f920cc39bc6d3b5621e9a86ba36a106e3d541aabd409ee7d70c231",
+    "output": "9d205b57ab166d23c81927899f99b6ccece261c679c96979c58fbde975637776",
+    "result": "5297c6c91a002e4ea712bb3ccb616944efc55f03a3cef381eafe85a12b439e3f",
+    "semantic": "a8f74965bc7c20d7b8de488e05b6c90e85caaab94b480adb71038165541a7c74",
+    "trajectory": "609a6b11bc5efdee6761986cbefa9b0a92db03a84e6d4fef6d842356ee5712c5",
+    "treatment_use": "be48c00d06af98783f37814cd5ac68233447d19cb1de8ad0f36fabfea13162c5"
+  },
+  "status": "fail",
+  "treatment_status": "pass"
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-environment-receipt.json b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-environment-receipt.json
new file mode 100644
index 000000000..f9e9e7619
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-environment-receipt.json
@@ -0,0 +1,5 @@
+{
+  "error": "environment receipt missing",
+  "schema_version": "deltawire-environment-receipt.v1",
+  "status": "fail"
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-plan-contract-receipt.json b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-plan-contract-receipt.json
new file mode 100644
index 000000000..bebb14ed9
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-plan-contract-receipt.json
@@ -0,0 +1,39 @@
+{
+  "checks": {
+    "exact_count_assertion": true,
+    "output_format": true,
+    "output_path": true,
+    "projected_records": true,
+    "schema_exists": true,
+    "schema_path": true,
+    "schema_sha256": true
+  },
+  "contract_sha256": "d3e1b7b3e5a4576540b3a2969d59f49818c7c0a12197b26d271e7e277b0054e5",
+  "inspect": {
+    "assertion_summary": [
+      "exact count 500"
+    ],
+    "cmd_check": "deltawire check .deltawire/plans/range-large.dw.json",
+    "cmd_render": "deltawire render .deltawire/plans/range-large.dw.json",
+    "cmd_validate": "deltawire validate .deltawire/plans/range-large.dw.json",
+    "cold_amp_ratio": "6892/845",
+    "output_format": "ndjson",
+    "output_path": "testdata/generated/output.ndjson",
+    "plan_id": "",
+    "plan_only_amp_ratio": "6892/559",
+    "plan_path": ".deltawire/plans/range-large.dw.json",
+    "plan_source_bytes": 559,
+    "primitives_used": [
+      "matrix",
+      "ranges"
+    ],
+    "projected_output_bytes": 6892,
+    "projected_records": 500,
+    "schema_path": ".deltawire/schemas/range-large.schema.json",
+    "schema_source_bytes": 286
+  },
+  "inspect_exit_code": 0,
+  "plan_sha256": "b1dcd5d61787bbc1a1873ae213b1105e7ac19184191eec599f5edd2a443404cb",
+  "schema_version": "plan-contract-receipt.v1",
+  "status": "pass"
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-semantic-result.json b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-semantic-result.json
new file mode 100644
index 000000000..f29151d44
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-semantic-result.json
@@ -0,0 +1,9 @@
+{
+  "error": null,
+  "exact_match": 1,
+  "oracle_result": {
+    "exact_match": 1
+  },
+  "schema_version": "semantic-result.v4",
+  "status": "pass"
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-treatment-use.json b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-treatment-use.json
new file mode 100644
index 000000000..bc8f8ec21
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/D1-treatment-use.json
@@ -0,0 +1,383 @@
+{
+  "arm": "D1",
+  "artifact_hashes": {
+    "config": "673b9cf053e3ba2263bfbc49034360ed8dba8d959bd167f8ff4b1f7cb4302b6f",
+    "in_container_contract": "9eb93bd1efb49fe95486d81529ff01ab6e2f20500eebd9146063380dea235f5a",
+    "output": "9d205b57ab166d23c81927899f99b6ccece261c679c96979c58fbde975637776",
+    "plan": "b1dcd5d61787bbc1a1873ae213b1105e7ac19184191eec599f5edd2a443404cb",
+    "schema": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5",
+    "state": "4c6bcc092e261346d8dd7ddcbc6a3cfc54e9908f08558840af43ba62dee1319d"
+  },
+  "checks": {
+    "check_invoked": true,
+    "check_not_proven_failed": true,
+    "in_container_contract_pass": true,
+    "independent_contract_pass": true,
+    "public_contract_access": true,
+    "render_invoked": true,
+    "render_not_proven_failed": true,
+    "retained_config": true,
+    "retained_in_container_contract": true,
+    "retained_output": true,
+    "retained_plan": true,
+    "retained_schema": true,
+    "retained_schema_integrity": true,
+    "retained_state": true,
+    "semantic_exact_pass": true,
+    "structured_activation": true,
+    "structured_discovery": true,
+    "validate_invoked": true,
+    "validate_not_proven_failed": true
+  },
+  "classification": "used_successfully",
+  "contamination": false,
+  "lifecycle_invocations": [
+    {
+      "command": "validate",
+      "invocation": "deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    },
+    {
+      "command": "validate",
+      "invocation": "deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    },
+    {
+      "command": "validate",
+      "invocation": "deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    },
+    {
+      "command": "validate",
+      "invocation": "deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    },
+    {
+      "command": "validate",
+      "invocation": "deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    },
+    {
+      "command": "validate",
+      "invocation": "deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    },
+    {
+      "command": "render",
+      "invocation": "deltawire render --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    },
+    {
+      "command": "check",
+      "invocation": "deltawire check --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    },
+    {
+      "command": "validate",
+      "invocation": "deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    },
+    {
+      "command": "render",
+      "invocation": "deltawire render --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    },
+    {
+      "command": "check",
+      "invocation": "deltawire check --repo . .deltawire/plans/range-large.dw.json",
+      "shell_exit_code": null
+    }
+  ],
+  "schema_version": "treatment-use-receipt.v4",
+  "shell_observations": [
+    {
+      "command": "gemini skills list",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "cat deltawire-environment-expectations.json",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "deltawire --help",
+      "observable": true,
+      "shell_exit_code": 2,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "deltawire help",
+      "observable": true,
+      "shell_exit_code": 2,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "mkdir -p .deltawire/plans && echo '{}' > .deltawire/plans/range-large.dw.json && deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "strings /usr/local/bin/deltawire | grep -i \"version\" | head -n 20",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "echo '{\"version\": \"1.0\"}' > .deltawire/plans/range-large.dw.json && deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "echo '{\"version\": \"deltawire.v1\"}' > .deltawire/plans/range-large.dw.json && deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "echo '{\"version\": \"plan.v1\"}' > .deltawire/plans/range-large.dw.json && deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "cat $(which verify_plan_contract.py)",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "which verify_plan_contract.py",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "cat /usr/local/bin/verify_plan_contract.py",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "grep -a \"unsupported plan version\" /usr/local/bin/deltawire",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "grep -a -o -E '.{0,10}version.{0,20}' /usr/local/bin/deltawire | strings | sort | uniq | head -n 30",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "grep -a -o -E '.{0,10}version.{0,20}' /usr/local/bin/deltawire | tr -cd '\\11\\12\\15\\40-\\176' | sort | uniq | head -n 30",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "echo '{\"version\": \"deltawire.plan.v1\"}' > .deltawire/plans/range-large.dw.json && deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output_path\": \"testdata/generated/output.ndjson\",\n  \"output_format\": \"ndjson\"\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json\n",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "grep -a -o -E 'json:\"[^\"]+\"' /usr/local/bin/deltawire | sort | uniq -c | sort -nr | head -n 50",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"range\",\n      \"start\": 1,\n      \"end_exclusive\": 501,\n      \"step\": 1\n    }\n  ]\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json\n",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"range\": {\n        \"start\": 1,\n        \"end_exclusive\": 501,\n        \"step\": 1\n      }\n    }\n  ]\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "grep -a -B 2 -A 2 -E 'json:\"(range|start|end_exclusive)\"' /usr/local/bin/deltawire",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "strings /usr/local/bin/deltawire | grep -o -E 'json:\"[^\"]+\"' | grep -v 'omitempty' | sort | uniq -c",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "cat /usr/local/bin/deltawire | tr -cd '\\11\\12\\15\\40-\\176' | grep -o -E 'json:\"[^\"]+\"' | sort | uniq -c",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"dimensions\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"range\",\n      \"start\": 1,\n      \"end_exclusive\": 501,\n      \"step\": 1\n    }\n  ]\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"range\"\n    }\n  ]\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 1,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "grep -a -o -E 'unknown generator kind: [^\"]*' /usr/local/bin/deltawire",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"cartesian\",\n      \"dimensions\": [\n        {\n          \"name\": \"index\",\n          \"start\": 1,\n          \"end_exclusive\": 501,\n          \"step\": 1\n        }\n      ]\n    }\n  ]\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"cartesian\",\n      \"dimensions\": [\n        {\n          \"name\": \"index\",\n          \"range\": {\n            \"start\": 1,\n            \"end_exclusive\": 501,\n            \"step\": 1\n          }\n        }\n      ]\n    }\n  ]\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 1,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "strings /usr/local/bin/deltawire | grep -i \"generator kind\" | head -n 10",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "cat /usr/local/bin/deltawire | tr -cd '\\11\\12\\15\\40-\\176' | grep -a -o -E '.{0,20}generator kind.{0,20}' | sort | uniq",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"product\",\n      \"dimensions\": [\n        {\n          \"name\": \"index\",\n          \"range\": {\n            \"start\": 1,\n            \"end_exclusive\": 501,\n            \"step\": 1\n          }\n        }\n      ]\n    }\n  ]\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 1,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "for kind in cartesian_product matrix grid rows repeated_rows combinations repeat sequence; do\ncat << EOF > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"\\$kind\",\n      \"dimensions\": [\n        {\n          \"name\": \"index\",\n          \"range\": {\n            \"start\": 1,\n            \"end_exclusive\": 501,\n            \"step\": 1\n          }\n        }\n      ]\n    }\n  ]\n}\nEOF\necho \"Trying \\$kind:\"\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json || true\ndone\n",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "for kind in cartesian cartesian_product product matrix grid rows repeated_rows repeat variants generator sequence set list; do\ncat << EOF > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"$kind\",\n      \"dimensions\": [\n        {\n          \"name\": \"index\",\n          \"range\": {\n            \"start\": 1,\n            \"end_exclusive\": 501,\n            \"step\": 1\n          }\n        }\n      ]\n    }\n  ]\n}\nEOF\necho \"Trying $kind:\"\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json || true\ndone",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"assertions\": {\n    \"record_count\": 500\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"matrix\",\n      \"dimensions\": [\n        {\n          \"name\": \"index\",\n          \"range\": {\n            \"start\": 1,\n            \"end_exclusive\": 501,\n            \"step\": 1\n          }\n        }\n      ]\n    }\n  ]\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"assertions\": {\n    \"count\": 500\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"matrix\",\n      \"dimensions\": [\n        {\n          \"name\": \"index\",\n          \"range\": {\n            \"start\": 1,\n            \"end_exclusive\": 501,\n            \"step\": 1\n          }\n        }\n      ]\n    }\n  ]\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": 3,
+      "shell_exit_provenance": "explicit_output",
+      "tool_status": "success"
+    },
+    {
+      "command": "cat << 'EOF' > .deltawire/plans/range-large.dw.json\n{\n  \"version\": \"deltawire.plan.v1\",\n  \"record_schema\": \".deltawire/schemas/range-large.schema.json\",\n  \"output\": {\n    \"path\": \"testdata/generated/output.ndjson\",\n    \"format\": \"ndjson\"\n  },\n  \"assertions\": {\n    \"count\": 500\n  },\n  \"generators\": [\n    {\n      \"name\": \"index\",\n      \"kind\": \"matrix\",\n      \"dimensions\": [\n        {\n          \"name\": \"index\",\n          \"range\": {\n            \"start\": 1,\n            \"end_exclusive\": 501,\n            \"step\": 1\n          }\n        }\n      ],\n      \"record\": {\n        \"index\": \"${index}\"\n      }\n    }\n  ]\n}\nEOF\ndeltawire inspect --repo . --format json .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "deltawire validate --repo . .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "deltawire render --repo . .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "head -n 5 testdata/generated/output.ndjson",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "deltawire check --repo . .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    },
+    {
+      "command": "deltawire validate --repo . .deltawire/plans/range-large.dw.json && \\\nverify_plan_contract.py --contract /task-contract.json --repo . --plan .deltawire/plans/range-large.dw.json --receipt .deltawire/plan-contract-receipt.json && \\\ndeltawire render --repo . .deltawire/plans/range-large.dw.json && \\\ndeltawire check --repo . .deltawire/plans/range-large.dw.json",
+      "observable": true,
+      "shell_exit_code": null,
+      "shell_exit_provenance": null,
+      "tool_status": "success"
+    }
+  ],
+  "status": "pass",
+  "trajectory_sha256": "609a6b11bc5efdee6761986cbefa9b0a92db03a84e6d4fef6d842356ee5712c5"
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/redaction.json b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/redaction.json
new file mode 100644
index 000000000..e369f97e0
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/redaction.json
@@ -0,0 +1,14 @@
+{
+  "schema_version": "evidence-redaction.v1",
+  "credential_patterns_checked": [
+    "exact sourced Gemini credential bytes",
+    "Google API key",
+    "GEMINI_API_KEY assignment",
+    "GitHub token",
+    "private key block"
+  ],
+  "unredacted_credentials_found": 0,
+  "credential_transport": "private temporary 0600 env file removed after Harbor exited",
+  "run_ledger_assignment": "<private-env-file>",
+  "raw_evidence_committable": true
+}
diff --git a/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/report.md b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/report.md
new file mode 100644
index 000000000..5e44f3551
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/report.md
@@ -0,0 +1,31 @@
+# Range-large v4 Measurement-Recovery Probe
+
+This directory preserves exactly one D1-only, non-benchmark v4 probe executed
+on `tbench-c4d`. Harbor 0.20.0 and Gemini 3.1 Pro Preview completed without an
+exception in 261.440381 seconds. The independent semantic oracle passed
+(`exact_match=1`) for the canonical 500 integer `index` records.
+
+The structured trajectory and retained artifacts passed the treatment-use
+boundary. `gemini skills list` preceded successful `activate_skill`; the public
+contract was accessed; DeltaWire validate, render, and check were invoked; the
+config, schema, plan, state, in-container contract receipt, and output were
+retained; schema integrity passed; and the independent plan-contract recheck
+passed with a projected count of 500.
+
+The deterministic environment generator also ran in Harbor's verifier phase
+and printed a passing `deltawire-environment-receipt.v1`: the binary was the
+expected executable, non-symlink regular file at `/usr/local/bin/deltawire`;
+its realpath, version, exit status, and SHA-256 passed; and the config, schema,
+and public-contract hashes passed. Harbor then attempted to copy
+`.deltawire/environment-receipt.json` from the main task container, where the
+file did not exist. The independently retained environment receipt was
+therefore missing, so the fail-closed end-to-end receipt recorded
+`environment_status=fail` and `status=fail`.
+
+No `probe-release.v2` was created and the authorized `range-large/r1` pair was
+not launched. Reported probe diagnostics are 934,798 input tokens, 10,344
+output tokens, 750,963 cache tokens, and 261.440381 seconds wall time. They are
+not a benchmark comparison, savings claim, or cost claim.
+
+`READY_FOR_72` remains false. Under the v4 one-shot stop condition, this probe
+will not be repaired or rerun and no v5 is implied or authorized.
diff --git a/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/run-ledger.json b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/run-ledger.json
new file mode 100644
index 000000000..72b3e3e2c
--- /dev/null
+++ b/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/run-ledger.json
@@ -0,0 +1,29 @@
+{
+  "mode": "probe-v4",
+  "ready_for_72": false,
+  "runs": [
+    {
+      "arm": "D1",
+      "cache_tokens": 750963,
+      "harbor_exit_code": 0,
+      "input_tokens": 934798,
+      "output_path": "labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/raw/D1/2026-07-22__13-20-47/range-large-v4__X4J4zSn/artifacts/testdata/generated/output.ndjson",
+      "output_sha256": "9d205b57ab166d23c81927899f99b6ccece261c679c96979c58fbde975637776",
+      "output_tokens": 10344,
+      "pair_id": "probe/range-large-v4",
+      "redacted_command": "uvx --from harbor==0.20.0 harbor run -a gemini-cli -m google/gemini-3.1-pro-preview -p /home/apple/deltawire-preflight/labs/20-deltawire/eval/probes/range-large-v4 -k 1 --jobs-dir /home/apple/deltawire-preflight/labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/raw/D1 --n-concurrent 1 --skill /home/apple/deltawire-preflight/labs/20-deltawire/eval/probes/range-large-v4/skill/deltawire --env-file '<private-env-file>'",
+      "repetition": 1,
+      "reported_model": "gemini-3.1-pro-preview",
+      "result_path": "labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/raw/D1/2026-07-22__13-20-47/range-large-v4__X4J4zSn/result.json",
+      "result_sha256": "5297c6c91a002e4ea712bb3ccb616944efc55f03a3cef381eafe85a12b439e3f",
+      "status": "completed",
+      "task": "range-large-v4",
+      "trajectory_path": "labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/raw/D1/2026-07-22__13-20-47/range-large-v4__X4J4zSn/agent/gemini-cli.trajectory.jsonl",
+      "trajectory_sha256": "609a6b11bc5efdee6761986cbefa9b0a92db03a84e6d4fef6d842356ee5712c5",
+      "trial_dir": "labs/20-deltawire/eval/results/preflight-v4/treatment-probe-range-large-v4/raw/D1/2026-07-22__13-20-47/range-large-v4__X4J4zSn",
+      "trial_id": "0b2fa7e9-2ad8-4789-90fb-bf635795eed9",
+      "wall_time_seconds": 261.440381
+    }
+  ],
+  "schema_version": "run-ledger.v4"
+}
diff --git a/labs/20-deltawire/eval/scripts/build_and_manifest.sh b/labs/20-deltawire/eval/scripts/build_and_manifest.sh
new file mode 100755
index 000000000..0515379e9
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/build_and_manifest.sh
@@ -0,0 +1,17 @@
+#!/usr/bin/env bash
+set -euo pipefail
+ROOT="$(cd "$(dirname "$0")/../../../.." && pwd)"
+LAB="$ROOT/labs/20-deltawire"
+GENERATED="$LAB/eval/.generated"
+mkdir -p "$GENERATED"
+(
+  cd "$LAB"
+  CGO_ENABLED=0 GOOS=linux GOARCH=amd64 GOWORK=off go build -trimpath -o "$GENERATED/deltawire-linux-amd64" ./cmd/deltawire
+)
+SHA="$(shasum -a 256 "$GENERATED/deltawire-linux-amd64" | awk '{print $1}')"
+for environment in "$LAB"/eval/tasks/*/environment "$LAB"/eval/probes/*/environment; do
+  mkdir -p "$environment/.generated"
+  cp "$GENERATED/deltawire-linux-amd64" "$environment/.generated/deltawire"
+  test "$(shasum -a 256 "$environment/.generated/deltawire" | awk '{print $1}')" = "$SHA"
+done
+printf '%s\n' "$SHA"
diff --git a/labs/20-deltawire/eval/scripts/compute_readiness.py b/labs/20-deltawire/eval/scripts/compute_readiness.py
new file mode 100644
index 000000000..85b562e1e
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/compute_readiness.py
@@ -0,0 +1,26 @@
+#!/usr/bin/env python3
+import json
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[4]
+BASE = ROOT / "labs/20-deltawire/eval/results/preflight-v2"
+
+
+def load(path): return json.loads(path.read_text(encoding="utf-8")) if path.is_file() else None
+
+
+def main():
+    probe = load(BASE / "treatment-probe-range-large-v2/D1-treatment-fidelity.json")
+    b0 = load(BASE / "range-large-r1/B0-treatment-fidelity.json")
+    d1 = load(BASE / "range-large-r1/D1-treatment-fidelity.json")
+    reasons = ["six-run canary remains incomplete", "final confirmatory manifest is not frozen", "explicit 72-run authorization is absent"]
+    probe_ok = bool(probe and probe.get("end_to_end_status") == "pass")
+    if not probe_ok: reasons.append("range-large treatment-fidelity probe did not pass all delivery and semantic gates")
+    if not b0 or not d1: reasons.append("range-large/r1 positive-control pair incomplete")
+    elif b0.get("classification") != "not_used" or d1.get("end_to_end_status") != "pass": reasons.append("range-large/r1 treatment separation not proven")
+    result = {"READY_FOR_72": False, "reasons": reasons, "probe_passed": probe_ok,
+              "positive_control_pair_passed": bool(b0 and d1 and b0.get("classification") == "not_used" and d1.get("end_to_end_status") == "pass")}
+    path = BASE / "readiness.json"; path.parent.mkdir(parents=True, exist_ok=True); path.write_text(json.dumps(result, indent=2)+"\n", encoding="utf-8")
+
+
+if __name__ == "__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/fidelity.py b/labs/20-deltawire/eval/scripts/fidelity.py
new file mode 100755
index 000000000..cd5d29800
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/fidelity.py
@@ -0,0 +1,82 @@
+#!/usr/bin/env python3
+import hashlib
+import json
+from pathlib import Path
+
+
+def sha256(path):
+    digest = hashlib.sha256()
+    with Path(path).open("rb") as handle:
+        for chunk in iter(lambda: handle.read(1024 * 1024), b""):
+            digest.update(chunk)
+    return digest.hexdigest()
+
+
+def tool_calls(trajectory):
+    calls = []
+    for line in Path(trajectory).read_text(encoding="utf-8", errors="replace").splitlines():
+        try: event = json.loads(line)
+        except json.JSONDecodeError: continue
+        for call in event.get("toolCalls", []):
+            outputs = []
+            for item in call.get("result", []):
+                response = item.get("functionResponse", {}).get("response", {})
+                outputs.append(str(response.get("output", "")))
+            calls.append({"name": call.get("name"), "args": call.get("args") or {},
+                          "status": call.get("status"), "output": "\n".join(outputs)})
+    return calls
+
+
+def shell_command(call):
+    if call["name"] not in {"run_shell_command", "shell", "execute"}: return ""
+    return str(call["args"].get("command") or call["args"].get("cmd") or "")
+
+
+def receipt(trajectory, trial_dir, expected_binary_sha, semantic_exact_match, arm="D1", applicability="expected", expected_schema_sha=None, expected_config_sha=None, expected_contract_sha=None, expected_binary_version="deltawire version dev"):
+    calls = tool_calls(trajectory)
+    shell = [(call, shell_command(call)) for call in calls if shell_command(call)]
+    skills_list_discovered = any("gemini skills list" in cmd and "deltawire" in call["output"].lower() and call["status"] == "success" for call, cmd in shell)
+    activated = any(call["name"] == "activate_skill" and str(call["args"].get("name") or call["args"].get("skill_name") or "").lower() == "deltawire" and call["status"] == "success" for call in calls)
+    discovered = skills_list_discovered or activated
+    binary_available = any("command -v deltawire" in cmd and "/usr/local/bin/deltawire" in call["output"] and call["status"] == "success" for call, cmd in shell)
+    binary_version_proven = any("deltawire version" in cmd and expected_binary_version in call["output"] and call["status"] == "success" for call, cmd in shell)
+    binary_hash_match = any(("sha256sum" in cmd or "shasum" in cmd) and expected_binary_sha in call["output"] and call["status"] == "success" for call, cmd in shell)
+    lifecycle = {}
+    for name in ("validate", "render", "check"):
+        lifecycle[name] = any(f"deltawire {name}" in cmd and call["status"] == "success" for call, cmd in shell)
+    trial = Path(trial_dir)
+    plans = sorted((trial / "artifacts").rglob("*.dw.json")) if (trial / "artifacts").is_dir() else []
+    plan = plans[0] if plans else None
+    prose = "\n".join(call["output"] for call in calls)
+    explicit_abstention = "too irregular for a compact deterministic plan" in prose.lower()
+    contract_access = any((("/task-contract.json" in shell_command(call) and any(x in shell_command(call) for x in ("cat ", "sed ", "python"))) or (call["name"] in {"read_file", "read_many_files"} and "/task-contract.json" in json.dumps(call["args"]))) and call["status"] == "success" for call in calls)
+    contract_hash_match = bool(expected_contract_sha) and any(("sha256sum" in cmd or "shasum" in cmd) and "/task-contract.json" in cmd and expected_contract_sha in call["output"] and call["status"] == "success" for call, cmd in shell)
+    schema_hash_match = bool(expected_schema_sha) and any(("sha256sum" in cmd or "shasum" in cmd) and expected_schema_sha in call["output"] and call["status"] == "success" for call, cmd in shell)
+    config_hash_proven = bool(expected_config_sha) and any(("sha256sum" in cmd or "shasum" in cmd) and ".deltawire/config.json" in cmd and expected_config_sha in call["output"] and call["status"] == "success" for call, cmd in shell)
+    contract_check = any("verify_plan_contract.py" in cmd and call["status"] == "success" for call, cmd in shell)
+    config_path = trial / "artifacts/.deltawire/config.json"
+    config_retained = config_path.is_file()
+    config_hash_match = config_retained and bool(expected_config_sha) and sha256(config_path) == expected_config_sha
+    deltawire_shell_use = any("deltawire" in cmd.lower() for _, cmd in shell)
+    used = any(lifecycle.values()) or activated or deltawire_shell_use or plan is not None
+    if arm == "B0": classification = "unexpected_use" if used else "not_used"
+    elif applicability == "should_abstain" and explicit_abstention: classification = "abstained"
+    elif all((skills_list_discovered, activated, binary_available, binary_version_proven, binary_hash_match, contract_access, contract_hash_match, schema_hash_match, config_hash_proven, config_hash_match, contract_check, *lifecycle.values(), plan is not None)): classification = "used_successfully"
+    elif used: classification = "attempted_failed"
+    else: classification = "not_used"
+    delivery_status = "pass" if classification == "used_successfully" else ("not_applicable" if arm == "B0" and classification == "not_used" else "fail")
+    semantic_status = "pass" if semantic_exact_match == 1 else "fail"
+    end_to_end_status = "pass" if semantic_status == "pass" and (delivery_status == "pass" or (arm == "B0" and classification == "not_used")) else "fail"
+    return {
+        "schema_version": "treatment-fidelity-receipt.v2", "classification": classification, "delivery_status": delivery_status, "semantic_status": semantic_status, "end_to_end_status": end_to_end_status,
+        "discovered": discovered, "skills_list_discovered": skills_list_discovered, "activated": activated, "binary_available": binary_available,
+        "binary_version": expected_binary_version if binary_version_proven else None, "binary_version_proven": binary_version_proven,
+        "binary_sha256": expected_binary_sha if binary_hash_match else None, "binary_hash_match": binary_hash_match,
+        "public_contract_accessed": contract_access, "public_contract_sha256": expected_contract_sha if contract_hash_match else None, "public_contract_hash_match": contract_hash_match,
+        "config_retained": config_retained, "config_sha256": sha256(config_path) if config_retained else None, "config_hash_proven": config_hash_proven, "config_hash_match": config_hash_match,
+        "schema_sha256": expected_schema_sha if schema_hash_match else None, "schema_hash_match": schema_hash_match, "contract_check_exit_0": contract_check,
+        "validate_exit_0": lifecycle["validate"], "render_exit_0": lifecycle["render"], "check_exit_0": lifecycle["check"],
+        "plan_path": str(plan.relative_to(trial)) if plan else None, "plan_sha256": sha256(plan) if plan else None,
+        "plan_bytes": plan.stat().st_size if plan else None, "semantic_exact_match": semantic_exact_match,
+        "trajectory_sha256": sha256(trajectory), "explicit_abstention": explicit_abstention, "deltawire_shell_use": deltawire_shell_use,
+    }
diff --git a/labs/20-deltawire/eval/scripts/generate_manifest.py b/labs/20-deltawire/eval/scripts/generate_manifest.py
new file mode 100644
index 000000000..ae2643447
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/generate_manifest.py
@@ -0,0 +1,49 @@
+#!/usr/bin/env python3
+import argparse
+import hashlib
+import json
+import subprocess
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[4]
+EVAL = ROOT / "labs/20-deltawire/eval"
+OUTPUT = EVAL / "manifests/frozen-72-manifest.json"
+
+
+def sha(path):
+    digest = hashlib.sha256()
+    with path.open("rb") as handle:
+        for chunk in iter(lambda: handle.read(1024 * 1024), b""): digest.update(chunk)
+    return digest.hexdigest()
+
+
+def main():
+    parser = argparse.ArgumentParser(); parser.add_argument("--input-commit", required=True); parser.add_argument("--binary-sha", required=True); args = parser.parse_args()
+    paths = [
+        EVAL / "skills/deltawire/SKILL.md", EVAL / "scripts/runner.py", EVAL / "scripts/fidelity.py",
+        EVAL / "scripts/semantic_oracle.py", EVAL / "analysis/extract_runs.py", EVAL / "analysis/analyze_paired.py",
+        EVAL / "matrices/72-run-matrix.json",
+    ]
+    paths += sorted((EVAL / "tasks").glob("*/task-spec.json"))
+    paths += sorted((EVAL / "tasks").glob("*/instruction.md"))
+    paths += sorted((EVAL / "tasks").glob("*/environment/Dockerfile"))
+    hashes = {str(path.relative_to(ROOT)): sha(path) for path in paths}
+    manifest = {
+        "schema_version": "deltawire-experiment-manifest.v2", "repository_commit": args.input_commit,
+        "deltawire_binary_sha256": args.binary_sha, "model": "google/gemini-3.1-pro-preview",
+        "expected_reported_model": "gemini-3.1-pro-preview", "agent": "gemini-cli",
+        "harbor_package": "harbor==0.20.0", "harbor_version": "0.20.0",
+        "harbor_command": ["uvx", "--from", "harbor==0.20.0", "harbor"],
+        "exact_tasks": sorted(path.parent.name for path in (EVAL / "tasks").glob("*/task-spec.json")),
+        "exact_arms": ["B0", "D1"], "exact_repetitions": 3, "exact_expected_run_count": 72,
+        "matrix_seed": 42, "input_hashes": hashes,
+        "token_field_definitions": {"input_tokens": "/agent_result/n_input_tokens", "output_tokens": "/agent_result/n_output_tokens", "cache_tokens": "/agent_result/n_cache_tokens"},
+        "correctness_definition": "Contract-driven semantic oracle exact_match=1.",
+        "failure_and_rerun_policy": "Stop on infrastructure or fidelity failure; no automatic reruns.",
+        "analysis_version": "v2.0.0",
+    }
+    OUTPUT.parent.mkdir(parents=True, exist_ok=True); OUTPUT.write_text(json.dumps(manifest, indent=2)+"\n", encoding="utf-8")
+    print(OUTPUT.relative_to(ROOT))
+
+
+if __name__ == "__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/generate_matrix.py b/labs/20-deltawire/eval/scripts/generate_matrix.py
new file mode 100644
index 000000000..d8115394a
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/generate_matrix.py
@@ -0,0 +1,21 @@
+#!/usr/bin/env python3
+import json
+import random
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[4]
+EVAL = ROOT / "labs/20-deltawire/eval"
+
+
+def main():
+    tasks = sorted(path.parent.name for path in (EVAL / "tasks").glob("*/task-spec.json"))
+    pairs = [{"pair_id": f"{task}/r{rep}", "task": task, "repetition": rep} for rep in (1, 2, 3) for task in tasks]
+    rng = random.Random(42); rng.shuffle(pairs); matrix = []
+    for pair in pairs:
+        arms = ["B0", "D1"] if rng.choice((True, False)) else ["D1", "B0"]
+        matrix.extend([{**pair, "arm": arm} for arm in arms])
+    assert len(matrix) == 72 and len({(r["pair_id"], r["arm"]) for r in matrix}) == 72
+    output = EVAL / "matrices/72-run-matrix.json"; output.write_text(json.dumps(matrix, indent=2)+"\n", encoding="utf-8")
+
+
+if __name__ == "__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/generate_preflight_v2_manifest.py b/labs/20-deltawire/eval/scripts/generate_preflight_v2_manifest.py
new file mode 100644
index 000000000..b554b2c48
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/generate_preflight_v2_manifest.py
@@ -0,0 +1,12 @@
+#!/usr/bin/env python3
+import argparse,hashlib,json
+from pathlib import Path
+ROOT=Path(__file__).resolve().parents[4]; EVAL=ROOT/"labs/20-deltawire/eval"; OUT=EVAL/"manifests/preflight-v2-range-large.json"
+def sha(p): return hashlib.sha256(p.read_bytes()).hexdigest()
+def main():
+ p=argparse.ArgumentParser(); p.add_argument("--input-commit",required=True); p.add_argument("--binary-sha",required=True); a=p.parse_args()
+ paths=[EVAL/"skills/deltawire/SKILL.md",EVAL/"scripts/runner.py",EVAL/"scripts/fidelity.py",EVAL/"scripts/verify_plan_contract.py",EVAL/"scripts/semantic_oracle.py",EVAL/"matrices/72-run-matrix.json"]
+ for base in (EVAL/"tasks/range-large",EVAL/"probes/range-large-v2"): paths += [x for x in base.rglob("*") if x.is_file() and ".generated" not in x.parts]
+ value={"schema_version":"deltawire-preflight-v2-manifest.v1","repository_commit":a.input_commit,"scope":["probe/range-large-v2","range-large/r1"],"deltawire_binary_sha256":a.binary_sha,"model":"google/gemini-3.1-pro-preview","expected_reported_model":"gemini-3.1-pro-preview","agent":"gemini-cli","harbor_package":"harbor==0.20.0","harbor_version":"0.20.0","harbor_command":["uvx","--from","harbor==0.20.0","harbor"],"input_hashes":{str(x.relative_to(ROOT)):sha(x) for x in sorted(set(paths))},"failure_policy":"Stop on any failed gate; never rerun v1 or launch the 72-run matrix."}
+ OUT.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n"); print(OUT.relative_to(ROOT))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/generate_preflight_v3_manifest.py b/labs/20-deltawire/eval/scripts/generate_preflight_v3_manifest.py
new file mode 100644
index 000000000..4984381b5
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/generate_preflight_v3_manifest.py
@@ -0,0 +1,58 @@
+#!/usr/bin/env python3
+"""Freeze the complete v3 probe and conditional-pair execution closure."""
+import argparse
+import hashlib
+import json
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[4]
+EVAL = ROOT / "labs/20-deltawire/eval"
+OUT = EVAL / "manifests/preflight-v3-range-large.json"
+
+
+def sha(path):
+    return hashlib.sha256(path.read_bytes()).hexdigest()
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--input-commit", required=True)
+    parser.add_argument("--binary-sha", required=True)
+    args = parser.parse_args()
+    paths = [
+        EVAL / "skills/deltawire/SKILL.md",
+        EVAL / "scripts/runner.py",
+        EVAL / "scripts/fidelity.py",
+        EVAL / "scripts/verify_plan_contract.py",
+        EVAL / "scripts/semantic_oracle.py",
+        EVAL / "scripts/task_contracts.py",
+        EVAL / "matrices/72-run-matrix.json",
+        EVAL / "manifests/preflight-v2-evidence-lock.json",
+    ]
+    for base in (EVAL / "tasks/range-large", EVAL / "probes/range-large-v3"):
+        paths += [item for item in base.rglob("*") if item.is_file() and ".generated" not in item.parts]
+    missing = [str(path.relative_to(ROOT)) for path in paths if not path.is_file()]
+    if missing:
+        raise SystemExit("missing v3 input: " + ", ".join(missing))
+    value = {
+        "schema_version": "deltawire-preflight-v3-manifest.v1",
+        "repository_commit": args.input_commit,
+        "scope": ["probe/range-large-v3", "range-large/r1"],
+        "deltawire_binary_sha256": args.binary_sha,
+        "deltawire_version": "deltawire version dev",
+        "model": "google/gemini-3.1-pro-preview",
+        "expected_reported_model": "gemini-3.1-pro-preview",
+        "agent": "gemini-cli",
+        "harbor_package": "harbor==0.20.0",
+        "harbor_version": "0.20.0",
+        "harbor_command": ["uvx", "--from", "harbor==0.20.0", "harbor"],
+        "input_hashes": {str(path.relative_to(ROOT)): sha(path) for path in sorted(set(paths))},
+        "failure_policy": "One v3 probe only; release one range-large/r1 pair only after raw-evidence rechecks; never run full.",
+        "ready_for_72": False,
+    }
+    OUT.write_text(json.dumps(value, indent=2, sort_keys=True) + "\n", encoding="utf-8")
+    print(OUT.relative_to(ROOT))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/scripts/generate_v1_evidence_lock.py b/labs/20-deltawire/eval/scripts/generate_v1_evidence_lock.py
new file mode 100644
index 000000000..3c25e460b
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/generate_v1_evidence_lock.py
@@ -0,0 +1,6 @@
+#!/usr/bin/env python3
+import hashlib,json
+from pathlib import Path
+ROOT=Path(__file__).resolve().parents[4]; BASE=ROOT/"labs/20-deltawire/eval/results/preflight-v1/treatment-probe-range-large"; OUT=BASE/"v1-evidence-lock.json"
+paths=[p for p in BASE.rglob("*") if p.is_file() and p.suffix != ".log" and p.name not in {OUT.name,"treatment-fidelity-receipt.json","report.md"}]
+OUT.write_text(json.dumps({"schema_version":"evidence-lock.v1","probe_version":"v1","immutable_files":{str(p.relative_to(ROOT)):hashlib.sha256(p.read_bytes()).hexdigest() for p in sorted(paths)}},indent=2,sort_keys=True)+"\n")
diff --git a/labs/20-deltawire/eval/scripts/generate_v2_evidence_lock.py b/labs/20-deltawire/eval/scripts/generate_v2_evidence_lock.py
new file mode 100644
index 000000000..6487915b5
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/generate_v2_evidence_lock.py
@@ -0,0 +1,50 @@
+#!/usr/bin/env python3
+"""Create the immutable, repository-relative lock for historical v2 evidence."""
+import hashlib
+import json
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[4]
+EVAL = ROOT / "labs/20-deltawire/eval"
+RESULT = EVAL / "results/preflight-v2/treatment-probe-range-large-v2"
+OUT = EVAL / "manifests/preflight-v2-evidence-lock.json"
+
+
+def one(pattern):
+    matches = sorted(RESULT.glob(pattern))
+    if len(matches) != 1:
+        raise SystemExit(f"v2 lock expected one {pattern!r}, found {len(matches)}")
+    return matches[0]
+
+
+def main():
+    files = [
+        EVAL / "manifests/preflight-v2-range-large.json",
+        RESULT / "run-ledger.json",
+        RESULT / "D1-treatment-fidelity.json",
+        RESULT / "D1-plan-contract-receipt.json",
+        RESULT / "report.md",
+        one("raw/D1/*/result.json"),
+        one("raw/D1/*/*/result.json"),
+        one("raw/D1/*/*/agent/gemini-cli.trajectory.jsonl"),
+        one("raw/D1/*/*/artifacts/testdata/generated/output.ndjson"),
+        one("raw/D1/*/*/artifacts/.deltawire/plans/range-large.dw.json"),
+        one("raw/D1/*/*/artifacts/.deltawire/plan-contract-receipt.json"),
+    ]
+    missing = [str(path.relative_to(ROOT)) for path in files if not path.is_file()]
+    if missing:
+        raise SystemExit("missing v2 evidence: " + ", ".join(missing))
+    value = {
+        "schema_version": "evidence-lock.v1",
+        "evidence": "range-large-v2-failed-treatment-probe",
+        "immutable_files": {
+            str(path.relative_to(ROOT)): hashlib.sha256(path.read_bytes()).hexdigest()
+            for path in files
+        },
+    }
+    OUT.write_text(json.dumps(value, indent=2, sort_keys=True) + "\n", encoding="utf-8")
+    print(OUT.relative_to(ROOT))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/scripts/patch_tests.py b/labs/20-deltawire/eval/scripts/patch_tests.py
new file mode 100644
index 000000000..69e0d5a47
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/patch_tests.py
@@ -0,0 +1,29 @@
+import os
+import glob
+
+base_dir = "labs/20-deltawire/eval/tasks"
+tasks = [
+    "irregular-large", "irregular-medium",
+    "matrix-large", "matrix-medium",
+    "mixed-large", "mixed-medium",
+    "range-large", "range-medium",
+    "rows-large", "rows-medium",
+    "variants-large", "variants-medium"
+]
+
+test_content_template = """#!/bin/bash
+OUTPUT_FILE=$(ls testdata/generated/*.ndjson 2>/dev/null | head -n 1)
+if [ -z "$OUTPUT_FILE" ]; then
+    echo '{{"exact_match": 0, "reason": "No output file generated"}}'
+    exit 0
+fi
+python3 ../../scripts/semantic_oracle.py "{task}" "$OUTPUT_FILE"
+"""
+
+for task in tasks:
+    test_path = os.path.join(base_dir, task, "tests", "test.sh")
+    with open(test_path, "w") as f:
+        f.write(test_content_template.format(task=task))
+    os.chmod(test_path, 0o755)
+
+print("Updated test.sh for all tasks to invoke semantic_oracle.py")
diff --git a/labs/20-deltawire/eval/scripts/run_paired.sh b/labs/20-deltawire/eval/scripts/run_paired.sh
old mode 100755
new mode 100644
index 2c6bd0918..01f0c5b0d
--- a/labs/20-deltawire/eval/scripts/run_paired.sh
+++ b/labs/20-deltawire/eval/scripts/run_paired.sh
@@ -1,2 +1,10 @@
-echo 'run_paired.sh stub'
-exit 0
\ No newline at end of file
+#!/usr/bin/env bash
+set -euo pipefail
+if [ "$#" -lt 2 ]; then
+  echo "usage: $0 <probe|pair|full> <manifest-sha> [runner arguments...]" >&2
+  exit 2
+fi
+MODE="$1"; SHA="$2"; shift 2
+python3 labs/20-deltawire/eval/scripts/runner.py "$MODE" \
+  --manifest labs/20-deltawire/eval/manifests/frozen-72-manifest.json \
+  --approve-manifest-sha "$SHA" "$@"
diff --git a/labs/20-deltawire/eval/scripts/run_pilot.sh b/labs/20-deltawire/eval/scripts/run_pilot.sh
deleted file mode 100755
index e3c4c4e95..000000000
--- a/labs/20-deltawire/eval/scripts/run_pilot.sh
+++ /dev/null
@@ -1,2 +0,0 @@
-echo 'run_pilot.sh stub'
-exit 0
\ No newline at end of file
diff --git a/labs/20-deltawire/eval/scripts/run_qwen_pilot.sh b/labs/20-deltawire/eval/scripts/run_qwen_pilot.sh
deleted file mode 100755
index 52432a62c..000000000
--- a/labs/20-deltawire/eval/scripts/run_qwen_pilot.sh
+++ /dev/null
@@ -1,2 +0,0 @@
-echo 'run_qwen_pilot.sh stub'
-exit 0
\ No newline at end of file
diff --git a/labs/20-deltawire/eval/scripts/runner.py b/labs/20-deltawire/eval/scripts/runner.py
new file mode 100644
index 000000000..113595e93
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/runner.py
@@ -0,0 +1,478 @@
+#!/usr/bin/env python3
+"""Fail-closed DeltaWire preflight runner.
+
+Only one versioned v3 treatment probe and, after a content-addressed release,
+one frozen adjacent range-large/r1 pair are executable. Historical probes and
+the 72-run mode are deliberately non-runnable.
+"""
+import argparse
+import hashlib
+import json
+import os
+import shlex
+import subprocess
+import sys
+import tempfile
+from datetime import datetime
+from pathlib import Path
+
+from fidelity import receipt, sha256
+
+ROOT = Path(__file__).resolve().parents[4]
+EVAL = ROOT / "labs/20-deltawire/eval"
+SKILL = EVAL / "skills/deltawire"
+V3_ROOT = EVAL / "results/preflight-v3/treatment-probe-range-large-v3"
+PAIR_ROOT = EVAL / "results/preflight-v3/range-large-r1"
+V3_MANIFEST = EVAL / "manifests/preflight-v3-range-large.json"
+
+
+def load(path):
+    return json.loads(Path(path).read_text(encoding="utf-8"))
+
+
+def write(path, value):
+    path = Path(path)
+    path.parent.mkdir(parents=True, exist_ok=True)
+    path.write_text(json.dumps(value, indent=2, sort_keys=True) + "\n", encoding="utf-8")
+
+
+def relative(path):
+    return str(Path(path).resolve().relative_to(ROOT.resolve()))
+
+
+def verify_hash_lock(path):
+    lock = load(path)
+    for name, expected in lock.get("immutable_files", {}).items():
+        target = ROOT / name
+        if not target.is_file() or sha256(target) != expected:
+            raise RuntimeError(f"immutable evidence mismatch: {name}")
+
+
+def verify_manifest(manifest):
+    if manifest.get("schema_version") != "deltawire-preflight-v3-manifest.v1":
+        raise RuntimeError("not a v3 preflight manifest")
+    if manifest.get("scope") != ["probe/range-large-v3", "range-large/r1"]:
+        raise RuntimeError("v3 manifest scope mismatch")
+    if manifest.get("harbor_version") != "0.20.0":
+        raise RuntimeError("frozen Harbor version mismatch")
+    if manifest.get("model") != "google/gemini-3.1-pro-preview" or manifest.get("agent") != "gemini-cli":
+        raise RuntimeError("frozen model or agent mismatch")
+    if manifest.get("deltawire_version") != "deltawire version dev":
+        raise RuntimeError("frozen DeltaWire version mismatch")
+    actual = subprocess.check_output(manifest["harbor_command"] + ["--version"], cwd=ROOT, text=True).strip()
+    if actual != manifest["harbor_version"]:
+        raise RuntimeError(f"Harbor version mismatch: {actual}")
+    for name, expected in manifest["input_hashes"].items():
+        target = ROOT / name
+        if not target.is_file() or sha256(target) != expected:
+            raise RuntimeError(f"frozen input mismatch: {name}")
+    verify_hash_lock(EVAL / "manifests/preflight-v2-evidence-lock.json")
+
+
+def load_approved_manifest(path, approved_sha):
+    path = Path(path).resolve()
+    if path != V3_MANIFEST.resolve():
+        raise RuntimeError("only the dedicated v3 manifest is accepted")
+    if sha256(path) != approved_sha:
+        raise RuntimeError("manifest SHA mismatch")
+    manifest = load(path)
+    verify_manifest(manifest)
+    return manifest
+
+
+def verify_binary(task_paths, expected):
+    central = EVAL / ".generated/deltawire-linux-amd64"
+    if not central.is_file() or sha256(central) != expected:
+        raise RuntimeError("central staged binary hash mismatch")
+    for task in task_paths:
+        staged = task / "environment/.generated/deltawire"
+        if not staged.is_file() or sha256(staged) != expected:
+            raise RuntimeError(f"staged binary hash mismatch: {staged}")
+
+
+def refuse_nonempty(root, label):
+    if root.exists() and any(root.iterdir()):
+        raise RuntimeError(f"{label} result root already contains evidence; reruns are forbidden")
+
+
+def trial_dirs(jobs):
+    return {path.resolve() for path in jobs.glob("*/*__*") if path.is_dir()}
+
+
+def private_env_file():
+    key = os.environ.get("GEMINI_API_KEY", "").strip()
+    if not key:
+        raise RuntimeError("GEMINI_API_KEY is required")
+    handle = tempfile.NamedTemporaryFile(mode="w", prefix="deltawire-harbor-", suffix=".env", delete=False)
+    try:
+        handle.write("GEMINI_API_KEY=" + key.replace("\n", "") + "\n")
+        handle.close()
+        os.chmod(handle.name, 0o600)
+        return Path(handle.name)
+    except Exception:
+        Path(handle.name).unlink(missing_ok=True)
+        raise
+
+
+def run_one(manifest, run, task_path, jobs, dry_run=False):
+    command = manifest["harbor_command"] + [
+        "run", "-a", manifest["agent"], "-m", manifest["model"], "-p", str(task_path),
+        "-k", "1", "--jobs-dir", str(jobs), "--n-concurrent", "1",
+    ]
+    if run["arm"] == "D1":
+        command += ["--skill", str(SKILL)]
+    redacted = shlex.join(command + ["--env-file", "<private-env-file>"])
+    if dry_run:
+        print(redacted)
+        return {**run, "status": "dry_run", "redacted_command": redacted}
+    env_file = private_env_file()
+    before = trial_dirs(jobs)
+    jobs.mkdir(parents=True, exist_ok=True)
+    child_env = os.environ.copy()
+    child_env.pop("GEMINI_API_KEY", None)
+    try:
+        try:
+            completed = subprocess.run(command + ["--env-file", str(env_file)], cwd=ROOT, env=child_env, check=False)
+            return_code = completed.returncode
+            run_error = None
+        except OSError as error:
+            return_code = None
+            run_error = str(error)
+    finally:
+        env_file.unlink(missing_ok=True)
+    created = trial_dirs(jobs) - before
+    entry = {**run, "status": "completed" if return_code == 0 else "harbor_failed",
+             "harbor_exit_code": return_code, "redacted_command": redacted}
+    if run_error:
+        entry["harbor_error"] = run_error
+    if len(created) != 1:
+        entry["artifact_error"] = f"expected one new trial, found {len(created)}"
+        return entry
+    trial = created.pop()
+    entry["trial_dir"] = relative(trial)
+    candidates = {
+        "result_path": trial / "result.json",
+        "trajectory_path": trial / "agent/gemini-cli.trajectory.jsonl",
+    }
+    outputs = sorted((trial / "artifacts").rglob("output.ndjson"))
+    if outputs:
+        candidates["output_path"] = outputs[0]
+    for field, path in candidates.items():
+        if path.is_file():
+            entry[field] = relative(path)
+            entry[field.replace("_path", "_sha256")] = sha256(path)
+    required = ("result_path", "trajectory_path", "output_path")
+    missing = [field for field in required if field not in entry]
+    if missing:
+        entry["artifact_error"] = "missing Harbor artifacts: " + ", ".join(missing)
+    return entry
+
+
+def semantic(task, output):
+    spec = EVAL / f"tasks/{task}/task-spec.json"
+    run = subprocess.run(
+        ["python3", str(EVAL / "scripts/semantic_oracle.py"), str(spec), str(output)],
+        capture_output=True, text=True, check=False,
+    )
+    if run.returncode:
+        return {"schema_version": "semantic-result.v1", "task": task, "exact_match": 0,
+                "status": "fail", "error": run.stderr.strip() or run.stdout.strip()}
+    value = json.loads(run.stdout)
+    return {"schema_version": "semantic-result.v1", "task": task,
+            "exact_match": value.get("exact_match", 0), "status": "pass" if value.get("exact_match") == 1 else "fail",
+            "oracle_result": value}
+
+
+def independent_plan_check(task, trial, binary, destination):
+    trial = Path(trial)
+    plans = sorted((trial / "artifacts/.deltawire/plans").glob("*.dw.json"))
+    if len(plans) != 1:
+        value = {"schema_version": "plan-contract-receipt.v1", "status": "fail", "checks": {},
+                 "error": f"expected one retained plan, found {len(plans)}"}
+        write(destination, value)
+        return value
+    command = [
+        "python3", str(EVAL / "scripts/verify_plan_contract.py"),
+        "--contract", str(EVAL / f"tasks/{task}/task-spec.json"), "--plan", str(plans[0]),
+        "--repo", str(trial / "artifacts"), "--deltawire", str(binary), "--receipt", str(destination),
+    ]
+    completed = subprocess.run(command, cwd=ROOT, capture_output=True, text=True, check=False)
+    if not Path(destination).is_file():
+        write(destination, {"schema_version": "plan-contract-receipt.v1", "status": "fail", "checks": {},
+                            "error": completed.stderr.strip() or completed.stdout.strip() or "checker wrote no receipt"})
+    return load(destination)
+
+
+def seconds_between(started, finished):
+    try:
+        return (datetime.fromisoformat(finished.replace("Z", "+00:00")) - datetime.fromisoformat(started.replace("Z", "+00:00"))).total_seconds()
+    except (AttributeError, TypeError, ValueError):
+        return None
+
+
+def failure_receipt(entry, arm, exact, reason):
+    semantic_status = "pass" if exact == 1 else "fail"
+    return {
+        "schema_version": "treatment-fidelity-receipt.v3", "classification": "not_used" if arm == "B0" else "attempted_failed",
+        "delivery_status": "not_applicable" if arm == "B0" else "fail", "semantic_status": semantic_status,
+        "end_to_end_status": "fail", "failure_reasons": [reason], "trial_id": entry.get("trial_id"),
+        "result_sha256": entry.get("result_sha256"), "trajectory_sha256": entry.get("trajectory_sha256"),
+        "output_sha256": entry.get("output_sha256"), "skills_list_discovered": False, "activated": False,
+        "binary_available": False, "binary_version_proven": False, "binary_hash_match": False,
+        "public_contract_accessed": False, "public_contract_hash_match": False,
+        "config_retained": False, "config_hash_proven": False, "config_hash_match": False,
+        "schema_hash_match": False, "retained_schema_integrity": False, "contract_check_exit_0": False,
+        "in_container_contract_check": "fail", "independent_contract_check": "fail",
+        "validate_exit_0": False, "render_exit_0": False, "check_exit_0": False,
+        "plan_path": None, "plan_sha256": None, "plan_bytes": None, "semantic_exact_match": exact,
+        "input_tokens": entry.get("input_tokens"), "output_tokens": entry.get("output_tokens"),
+        "cache_tokens": entry.get("cache_tokens"), "wall_time_seconds": entry.get("wall_time_seconds"),
+        "harbor_exit_0": entry.get("harbor_exit_code") == 0, "artifact_set_complete": False,
+    }
+
+
+def finalize(manifest, entry, result_root):
+    arm = entry["arm"]
+    task = entry.get("contract_task", entry["task"])
+    semantic_path = result_root / f"{arm}-semantic-result.json"
+    output = ROOT / entry["output_path"] if entry.get("output_path") else None
+    semantic_result = semantic(task, output) if output and output.is_file() else {
+        "schema_version": "semantic-result.v1", "task": task, "exact_match": 0, "status": "fail", "error": "output missing"}
+    write(semantic_path, semantic_result)
+    exact = semantic_result["exact_match"]
+    result_path = ROOT / entry["result_path"] if entry.get("result_path") else None
+    trajectory = ROOT / entry["trajectory_path"] if entry.get("trajectory_path") else None
+    trial = ROOT / entry["trial_dir"] if entry.get("trial_dir") else None
+    result_data = load(result_path) if result_path and result_path.is_file() else {}
+    entry["trial_id"] = result_data.get("id") or result_data.get("trial_name")
+    entry["started_at"] = result_data.get("started_at")
+    entry["finished_at"] = result_data.get("finished_at")
+    entry["wall_time_seconds"] = seconds_between(entry["started_at"], entry["finished_at"])
+    agent_result = result_data.get("agent_result") or {}
+    for source, target in (("n_input_tokens", "input_tokens"), ("n_output_tokens", "output_tokens"), ("n_cache_tokens", "cache_tokens")):
+        entry[target] = agent_result.get(source)
+    model = ((result_data.get("agent_info") or {}).get("model_info") or {}).get("name")
+    entry["reported_model"] = model
+    contract = EVAL / f"tasks/{task}/task-spec.json"
+    spec = load(contract)
+    config_source = EVAL / f"tasks/{task}/environment/.deltawire/config.json"
+    independent_path = result_root / f"{arm}-plan-contract-receipt.json"
+    if trajectory and trajectory.is_file() and trial and trial.is_dir():
+        evidence = receipt(
+            trajectory, trial, manifest["deltawire_binary_sha256"], exact, arm,
+            spec["deltawire_applicability"], spec["authoritative_schema"]["sha256"],
+            sha256(config_source), sha256(contract), manifest["deltawire_version"],
+        )
+        if arm == "D1":
+            independent = independent_plan_check(task, trial, EVAL / ".generated/deltawire-linux-amd64", independent_path)
+            evidence["independent_contract_check"] = "pass" if independent.get("status") == "pass" else "fail"
+            evidence["independent_contract_receipt"] = relative(independent_path)
+            evidence["projected_record_count"] = (independent.get("inspect") or {}).get("projected_records")
+        else:
+            evidence["independent_contract_check"] = "not_applicable"
+        schema_path = trial / "artifacts" / spec["authoritative_schema"]["path"]
+        evidence["retained_schema_integrity"] = schema_path.is_file() and sha256(schema_path) == spec["authoritative_schema"]["sha256"]
+        in_container = trial / "artifacts/.deltawire/plan-contract-receipt.json"
+        evidence["in_container_contract_receipt"] = relative(in_container) if in_container.is_file() else None
+        evidence["in_container_contract_receipt_sha256"] = sha256(in_container) if in_container.is_file() else None
+        evidence["in_container_contract_check"] = "pass" if in_container.is_file() and load(in_container).get("status") == "pass" else "fail"
+        evidence["end_to_end_status"] = "pass" if evidence["semantic_status"] == "pass" and (
+            arm == "B0" and evidence["classification"] == "not_used" or
+            arm == "D1" and evidence["delivery_status"] == "pass" and evidence["independent_contract_check"] == "pass" and
+            evidence["in_container_contract_check"] == "pass" and evidence["retained_schema_integrity"]
+        ) else "fail"
+    else:
+        evidence = failure_receipt(entry, arm, exact, entry.get("artifact_error", "trajectory or trial missing"))
+        if arm == "D1":
+            write(independent_path, {"schema_version": "plan-contract-receipt.v1", "status": "fail", "checks": {}, "error": "trial artifacts unavailable"})
+            evidence["independent_contract_receipt"] = relative(independent_path)
+    evidence["schema_version"] = "treatment-fidelity-receipt.v3"
+    evidence["trial_id"] = entry.get("trial_id")
+    evidence["result_sha256"] = entry.get("result_sha256")
+    evidence["output_sha256"] = entry.get("output_sha256")
+    evidence["semantic_result_sha256"] = sha256(semantic_path)
+    evidence["input_tokens"] = entry.get("input_tokens")
+    evidence["output_tokens"] = entry.get("output_tokens")
+    evidence["cache_tokens"] = entry.get("cache_tokens")
+    evidence["started_at"] = entry.get("started_at")
+    evidence["finished_at"] = entry.get("finished_at")
+    evidence["wall_time_seconds"] = entry.get("wall_time_seconds")
+    evidence["reported_model"] = entry.get("reported_model")
+    evidence["model_match"] = model == manifest["expected_reported_model"]
+    evidence["harbor_exit_0"] = entry.get("harbor_exit_code") == 0
+    evidence["artifact_set_complete"] = not entry.get("artifact_error")
+    if not evidence["model_match"] or not evidence["harbor_exit_0"] or not evidence["artifact_set_complete"]:
+        evidence["end_to_end_status"] = "fail"
+    treatment_path = result_root / f"{arm}-treatment-fidelity.json"
+    write(treatment_path, evidence)
+    return evidence, semantic_path, treatment_path
+
+
+def release_payload(manifest_path, ledger_path, entry, semantic_path, treatment_path, independent_path):
+    trial = ROOT / entry["trial_dir"]
+    required = [
+        Path(manifest_path), Path(ledger_path), ROOT / entry["result_path"], ROOT / entry["trajectory_path"],
+        ROOT / entry["output_path"], trial / "artifacts/.deltawire/plans/range-large.dw.json",
+        trial / "artifacts/.deltawire/config.json", trial / "artifacts/.deltawire/schemas/range-large.schema.json",
+        trial / "artifacts/.deltawire/plan-contract-receipt.json", Path(independent_path), Path(semantic_path), Path(treatment_path),
+    ]
+    missing = [str(path) for path in required if not path.is_file()]
+    if missing:
+        raise RuntimeError("probe release artifacts missing: " + ", ".join(missing))
+    payload = {
+        "schema_version": "probe-release.v1", "probe": "range-large-v3", "status": "pass",
+        "manifest_sha256": sha256(manifest_path),
+        "evidence_hashes": {relative(path): sha256(path) for path in required},
+        "ready_for_72": False,
+    }
+    canonical = json.dumps(payload, sort_keys=True, separators=(",", ":")).encode()
+    payload["release_id"] = hashlib.sha256(canonical).hexdigest()
+    return payload
+
+
+def verify_probe_release(manifest_path):
+    release_path = V3_ROOT / "probe-release.json"
+    if not release_path.is_file():
+        raise RuntimeError("pair requires persisted probe-release.v1")
+    release = load(release_path)
+    release_id = release.pop("release_id", None)
+    expected_id = hashlib.sha256(json.dumps(release, sort_keys=True, separators=(",", ":")).encode()).hexdigest()
+    release["release_id"] = release_id
+    if release.get("schema_version") != "probe-release.v1" or release.get("status") != "pass" or release_id != expected_id:
+        raise RuntimeError("invalid or fabricated probe release")
+    if release.get("manifest_sha256") != sha256(manifest_path):
+        raise RuntimeError("probe release manifest mismatch")
+    for name, expected in release.get("evidence_hashes", {}).items():
+        target = ROOT / name
+        if not target.is_file() or sha256(target) != expected:
+            raise RuntimeError(f"probe release raw evidence mismatch: {name}")
+    treatment = load(V3_ROOT / "D1-treatment-fidelity.json")
+    if treatment.get("end_to_end_status") != "pass" or treatment.get("independent_contract_check") != "pass":
+        raise RuntimeError("probe release treatment evidence is not end-to-end pass")
+    ledger = load(V3_ROOT / "run-ledger.json")
+    if len(ledger.get("runs", [])) != 1:
+        raise RuntimeError("probe release ledger is not exactly one run")
+    entry = ledger["runs"][0]
+    raw_semantic = semantic("range-large", ROOT / entry["output_path"])
+    if raw_semantic.get("exact_match") != 1:
+        raise RuntimeError("probe release semantic recheck failed")
+    with tempfile.TemporaryDirectory() as temp:
+        checked = independent_plan_check("range-large", ROOT / entry["trial_dir"], EVAL / ".generated/deltawire-linux-amd64", Path(temp) / "receipt.json")
+        if checked.get("status") != "pass":
+            raise RuntimeError("probe release retained-artifact recheck failed")
+    rerendered = receipt(
+        ROOT / entry["trajectory_path"], ROOT / entry["trial_dir"], treatment["binary_sha256"], 1, "D1", "expected",
+        treatment["schema_sha256"], sha256(EVAL / "tasks/range-large/environment/.deltawire/config.json"),
+        sha256(EVAL / "tasks/range-large/task-spec.json"), "deltawire version dev",
+    )
+    if rerendered.get("delivery_status") != "pass":
+        raise RuntimeError("probe release structured trajectory recheck failed")
+    return release
+
+
+def frozen_pair():
+    matrix = load(EVAL / "matrices/72-run-matrix.json")
+    positions = [i for i, run in enumerate(matrix) if run.get("pair_id") == "range-large/r1"]
+    if len(positions) != 2 or positions[1] != positions[0] + 1:
+        raise RuntimeError("range-large/r1 is not one frozen adjacent pair")
+    runs = [matrix[i] for i in positions]
+    if [run.get("arm") for run in runs] != ["D1", "B0"]:
+        raise RuntimeError("frozen range-large/r1 order must be D1 then B0")
+    return runs
+
+
+def parse_args():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("mode", choices=("probe-v3", "pair", "full"))
+    parser.add_argument("--manifest", required=True)
+    parser.add_argument("--approve-manifest-sha", required=True)
+    parser.add_argument("--pair-id")
+    parser.add_argument("--allow-72", action="store_true")
+    parser.add_argument("--dry-run", action="store_true")
+    return parser.parse_args()
+
+
+def main():
+    args = parse_args()
+    if args.mode == "full":
+        raise RuntimeError("full mode is unconditionally unauthorized in this preflight runner")
+    manifest_path = Path(args.manifest)
+    manifest = load_approved_manifest(manifest_path, args.approve_manifest_sha)
+    if args.mode == "probe-v3":
+        runs = [{"pair_id": "probe/range-large-v3", "task": "range-large-treatment-probe-v3", "contract_task": "range-large", "arm": "D1", "repetition": 1}]
+        task_paths = [EVAL / "probes/range-large-v3"]
+        result_root = V3_ROOT
+        if not args.dry_run:
+            refuse_nonempty(result_root, "v3 probe")
+    else:
+        if args.pair_id != "range-large/r1":
+            raise RuntimeError("this preflight permits only range-large/r1")
+        verify_probe_release(manifest_path)
+        runs = frozen_pair()
+        task_paths = [EVAL / f"tasks/{run['task']}" for run in runs]
+        result_root = PAIR_ROOT
+        if not args.dry_run:
+            refuse_nonempty(result_root, "range-large/r1 pair")
+    verify_binary(task_paths, manifest["deltawire_binary_sha256"])
+    ledger_path = result_root / "run-ledger.json"
+    ledger = {"schema_version": "run-ledger.v3", "mode": args.mode, "ready_for_72": False, "runs": []}
+    failure = None
+    for run, task_path in zip(runs, task_paths):
+        entry = run_one(manifest, run, task_path, result_root / "raw" / run["arm"], args.dry_run)
+        ledger["runs"].append(entry)
+        if args.dry_run:
+            continue
+        write(ledger_path, ledger)
+        try:
+            evidence, semantic_path, treatment_path = finalize(manifest, entry, result_root)
+        except Exception as error:
+            semantic_path = result_root / f"{entry['arm']}-semantic-result.json"
+            write(semantic_path, {"schema_version": "semantic-result.v1", "task": entry.get("contract_task", entry["task"]),
+                                  "exact_match": 0, "status": "fail", "error": str(error)})
+            independent_path = result_root / f"{entry['arm']}-plan-contract-receipt.json"
+            if entry["arm"] == "D1":
+                write(independent_path, {"schema_version": "plan-contract-receipt.v1", "status": "fail", "checks": {}, "error": str(error)})
+            evidence = failure_receipt(entry, entry["arm"], 0, str(error))
+            evidence["semantic_result_sha256"] = sha256(semantic_path)
+            treatment_path = result_root / f"{entry['arm']}-treatment-fidelity.json"
+            write(treatment_path, evidence)
+        write(ledger_path, ledger)
+        if evidence["semantic_status"] != "pass":
+            failure = "independent semantic oracle failed"
+        elif args.mode == "probe-v3" and evidence["end_to_end_status"] != "pass":
+            failure = "treatment-fidelity v3 probe failed"
+        elif args.mode == "pair" and entry["arm"] == "D1" and evidence["end_to_end_status"] != "pass":
+            failure = "D1 treatment fidelity failed"
+        elif args.mode == "pair" and entry["arm"] == "B0" and evidence["classification"] != "not_used":
+            failure = "B0 baseline contamination"
+        if failure:
+            break
+        if args.mode == "probe-v3":
+            independent_path = result_root / "D1-plan-contract-receipt.json"
+            release = release_payload(manifest_path, ledger_path, entry, semantic_path, treatment_path, independent_path)
+            write(result_root / "probe-release.json", release)
+    if args.dry_run:
+        return
+    if failure:
+        raise RuntimeError(failure)
+    if args.mode == "pair":
+        if len(ledger["runs"]) != 2:
+            raise RuntimeError("pair did not complete both frozen arms")
+        d1, b0 = ledger["runs"]
+        summary = {
+            "schema_version": "paired-preflight-summary.v1", "pair_id": "range-large/r1", "order": ["D1", "B0"],
+            "ready_for_72": False, "claim": "comparison_only_no_savings_claim",
+            "D1": {key: d1.get(key) for key in ("trial_id", "input_tokens", "output_tokens", "cache_tokens", "wall_time_seconds")},
+            "B0": {key: b0.get(key) for key in ("trial_id", "input_tokens", "output_tokens", "cache_tokens", "wall_time_seconds")},
+            "delta_D1_minus_B0": {key: (d1.get(key) - b0.get(key)) if d1.get(key) is not None and b0.get(key) is not None else None for key in ("input_tokens", "output_tokens", "cache_tokens", "wall_time_seconds")},
+        }
+        write(result_root / "pair-summary.json", summary)
+    print(f"{args.mode} completed")
+
+
+if __name__ == "__main__":
+    try:
+        main()
+    except (OSError, ValueError, KeyError, RuntimeError, subprocess.CalledProcessError) as error:
+        print(f"ERROR: {error}", file=sys.stderr)
+        raise SystemExit(1)
diff --git a/labs/20-deltawire/eval/scripts/semantic_oracle.py b/labs/20-deltawire/eval/scripts/semantic_oracle.py
new file mode 100755
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/scripts/task_contracts.py b/labs/20-deltawire/eval/scripts/task_contracts.py
new file mode 100644
index 000000000..abc2cd1e6
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/task_contracts.py
@@ -0,0 +1,71 @@
+#!/usr/bin/env python3
+"""Generate and validate public/hidden projections for all 12 tasks and probe v3.
+
+The v2 probe is historical evidence and is deliberately never written by this
+module.  Its bytes are protected by the v2 evidence lock.
+"""
+import argparse, hashlib, json, shutil
+from pathlib import Path
+ROOT=Path(__file__).resolve().parents[4]; EVAL=ROOT/"labs/20-deltawire/eval"; TASKS=EVAL/"tasks"; PROBE_V2=EVAL/"probes/range-large-v2"; PROBE_V3=EVAL/"probes/range-large-v3"; ORACLE=EVAL/"scripts/semantic_oracle.py"; CHECKER=EVAL/"scripts/verify_plan_contract.py"; OUTPUT="testdata/generated/output.ndjson"
+CONFIG={"version":"deltawire.config.v1","plans_dir":".deltawire/plans","schemas_dir":".deltawire/schemas","state_file":".deltawire/state.json","limits":{"max_plan_bytes":1048576,"max_schema_bytes":1048576,"max_records":100000,"max_output_bytes":104857600}}
+DEFS={
+"matrix-medium":("matrix","medium",6,"optional",{"dimensions":{"role":["guest","member","admin"],"method":["GET","POST"]},"constants":{"status":"deny"},"field_order":["role","method","status"]}),
+"matrix-large":("matrix","large",48,"expected",{"dimensions":{"role":["guest","member","admin","superuser"],"method":["GET","POST","PUT","DELETE"],"endpoint":["/api","/web","/login"]},"constants":{"status":"deny"},"field_order":["role","method","endpoint","status"]}),
+"range-medium":("range","medium",50,"optional",{"start":1,"field":"index","field_order":["index"]}),"range-large":("range","large",500,"expected",{"start":1,"field":"index","field_order":["index"]}),
+"rows-medium":("rows","medium",10,"optional",{"start":1,"field_order":["id","name","active"],"name_template":"User {id}","constants":{"active":True}}),"rows-large":("rows","large",100,"expected",{"start":1,"field_order":["id","name","active"],"name_template":"User {id}","constants":{"active":True}}),
+"variants-medium":("variants","medium",6,"optional",{"base":{"id":1,"type":"A","value":100},"value_step":10,"field_order":["id","type","value"]}),"variants-large":("variants","large",51,"expected",{"base":{"id":1,"type":"A","value":100},"value_step":10,"field_order":["id","type","value"]}),
+"mixed-medium":("mixed","medium",25,"optional",{"standard_count":20,"exception_count":5,"error_code":500,"field_order":["id","type"],"exception_field_order":["id","type","error_code"]}),"mixed-large":("mixed","large",250,"expected",{"standard_count":200,"exception_count":50,"error_code":500,"field_order":["id","type"],"exception_field_order":["id","type","error_code"]}),
+"irregular-medium":("irregular","medium",15,"should_abstain",{"id_start":1,"dynamic_field_prefix":"field_","field_order":["id","dynamic"]}),"irregular-large":("irregular","large",150,"should_abstain",{"id_start":1,"dynamic_field_prefix":"field_","field_order":["id","dynamic"]})}
+def canonical(v): return json.dumps(v,indent=2,sort_keys=True)+"\n"
+def schema(name):
+ f,_,n,_,g=DEFS[name]
+ if f=="matrix": props={**{k:{"enum":v} for k,v in g["dimensions"].items()},**{k:{"const":v} for k,v in g["constants"].items()}}; required=g["field_order"]
+ elif f=="range": props={g["field"]:{"type":"integer","minimum":g["start"],"maximum":g["start"]+n-1}}; required=g["field_order"]
+ elif f=="rows": props={"id":{"type":"integer","minimum":g["start"],"maximum":g["start"]+n-1},"name":{"type":"string","pattern":"^User [0-9]+$"},"active":{"const":True}}; required=g["field_order"]
+ elif f=="variants": props={"id":{"const":g["base"]["id"]},"type":{"const":g["base"]["type"]},"value":{"enum":[g["base"]["value"]+i*g["value_step"] for i in range(n)]}}; required=g["field_order"]
+ elif f=="mixed":
+  std={"type":"object","properties":{"id":{"type":"integer"},"type":{"const":"std"}},"required":["id","type"],"additionalProperties":False}; exc={"type":"object","properties":{"id":{"type":"integer"},"type":{"const":"exc"},"error_code":{"const":g["error_code"]}},"required":["id","type","error_code"],"additionalProperties":False}; return {"$schema":"https://json-schema.org/draft/2020-12/schema","title":name,"anyOf":[std,exc]}
+ else: return {"$schema":"https://json-schema.org/draft/2020-12/schema","title":name,"type":"object","properties":{"id":{"type":"integer","minimum":g["id_start"],"maximum":g["id_start"]+n-1}},"patternProperties":{f"^{g['dynamic_field_prefix']}[0-9]+$":{"type":"integer"}},"required":["id"],"minProperties":2,"maxProperties":2,"additionalProperties":False}
+ return {"$schema":"https://json-schema.org/draft/2020-12/schema","title":name,"type":"object","properties":props,"required":required,"additionalProperties":False}
+def contract(name):
+ f,size,n,app,g=DEFS[name]; s=canonical(schema(name)).encode(); return {"schema_version":"task-spec.v1","task":name,"family":f,"size":size,"output":{"path":OUTPUT,"format":"ndjson"},"record_count":n,"ordering":"canonical_generation_order","deltawire_applicability":app,"authoritative_schema":{"path":f".deltawire/schemas/{name}.schema.json","sha256":hashlib.sha256(s).hexdigest()},"generation":g}
+def instruction(s): return f"Generate the exact `{s['task']}` dataset defined by public `/task-contract.json`. Write exactly {s['record_count']} JSON records in canonical order as NDJSON to `{OUTPUT}`. Every record, field, value, and ordering rule must match this complete embedded contract:\n\n```json\n"+canonical(s)+"```\n"
+def probe_instruction(s): return "This is the versioned range-large-v3 non-benchmark DeltaWire treatment-fidelity recovery probe. Before activation, run `gemini skills list` and retain its structured result. Then activate the `deltawire` skill using the structured tool and follow it exactly. Read public `/task-contract.json`; prove path, version, and SHA-256 of `/usr/local/bin/deltawire`; prove the immutable public contract, config, and schema hashes. Create only `.deltawire/plans/range-large.dw.json`, then run validate, `verify_plan_contract.py`, render, and check. Do not edit the config or schema and do not generate expanded records manually. Complete contract:\n\n```json\n"+canonical(s)+"```\n"
+def write(path,v): path.parent.mkdir(parents=True,exist_ok=True); path.write_text(canonical(v))
+DOCKER="""FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
+"""
+def project(base,s,text):
+ for p in (base/"task-spec.json",base/"tests/task-spec.json",base/"environment/task-contract.json"): write(p,s)
+ for p in (base/"authoritative-schema.json",base/"environment"/s["authoritative_schema"]["path"]): write(p,schema(s["task"]))
+ write(base/"environment/.deltawire/config.json",CONFIG); (base/"instruction.md").write_text(text); shutil.copyfile(ORACLE,base/"tests/semantic_oracle.py"); shutil.copyfile(CHECKER,base/"environment/verify_plan_contract.py"); (base/"environment/Dockerfile").write_text(DOCKER)
+ if base.joinpath("task.toml").is_file():
+  artifact_line=f'artifacts = ["{OUTPUT}", ".deltawire/config.json", ".deltawire/plans/{s["task"]}.dw.json", "{s["authoritative_schema"]["path"]}", ".deltawire/plan-contract-receipt.json"]'
+  lines=base.joinpath("task.toml").read_text().splitlines(); lines=[artifact_line if line.startswith("artifacts =") else line for line in lines]; base.joinpath("task.toml").write_text("\n".join(lines)+"\n")
+def sync():
+ for name in sorted(DEFS): project(TASKS/name,contract(name),instruction(contract(name)))
+ s=contract("range-large"); project(PROBE_V3,s,probe_instruction(s)); shutil.copyfile(TASKS/"range-large/tests/test.sh",PROBE_V3/"tests/test.sh"); (PROBE_V3/"task.toml").write_text('''schema_version = "1.3"\nartifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/range-large.dw.json", ".deltawire/schemas/range-large.schema.json", ".deltawire/plan-contract-receipt.json"]\n\n[task]\nname = "operatorstack/range-large-treatment-probe-v3"\ndescription = "Versioned non-benchmark external-contract recovery probe"\nauthors = []\nkeywords = []\n\n[metadata]\nbenchmark_result = false\nprobe_version = "v3"\n\n[verifier]\ntimeout_sec = 600.0\n\n[agent]\ntimeout_sec = 600.0\n\n[environment]\nnetwork_mode = "public"\nbuild_timeout_sec = 600.0\nos = "linux"\nmcp_servers = []\n''')
+def check():
+ errors=[]
+ for name in sorted(DEFS):
+  s=contract(name); base=TASKS/name
+  for p in (base/"task-spec.json",base/"tests/task-spec.json",base/"environment/task-contract.json"):
+   if not p.is_file() or p.read_text()!=canonical(s): errors.append(f"stale projection: {p.relative_to(ROOT)}")
+  text=base.joinpath("instruction.md").read_text()
+  if text!=instruction(s) or "tests/" in text: errors.append(f"invalid instruction: {name}")
+  if hashlib.sha256(base.joinpath("authoritative-schema.json").read_bytes()).hexdigest()!=s["authoritative_schema"]["sha256"]: errors.append(f"schema hash: {name}")
+  if base.joinpath("environment/verify_plan_contract.py").read_bytes()!=CHECKER.read_bytes(): errors.append(f"checker: {name}")
+ s=contract("range-large")
+ if PROBE_V3.joinpath("instruction.md").read_text()!=probe_instruction(s) or "tests/" in PROBE_V3.joinpath("instruction.md").read_text(): errors.append("invalid v3 probe")
+ if 'operatorstack/range-large-treatment-probe-v3' not in PROBE_V3.joinpath("task.toml").read_text(): errors.append("invalid v3 task name")
+ if PROBE_V2.resolve() == PROBE_V3.resolve(): errors.append("v2/v3 probe paths collide")
+ if errors: raise SystemExit("\n".join(errors))
+ print("Task contract consistency passed.")
+if __name__=="__main__":
+ p=argparse.ArgumentParser(); p.add_argument("mode",choices=("sync","check")); a=p.parse_args(); sync() if a.mode=="sync" else check()
diff --git a/labs/20-deltawire/eval/scripts/test_fidelity.py b/labs/20-deltawire/eval/scripts/test_fidelity.py
new file mode 100755
index 000000000..9f03dbd84
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/test_fidelity.py
@@ -0,0 +1,71 @@
+#!/usr/bin/env python3
+import json
+import tempfile
+from pathlib import Path
+from fidelity import receipt
+
+SHA = "a" * 64
+SCHEMA_SHA = "b" * 64
+CONTRACT_SHA = "c" * 64
+
+
+def event(name, args, output="", status="success"):
+    return {"toolCalls": [{"name": name, "args": args, "status": status,
+                           "result": [{"functionResponse": {"response": {"output": output}}}]}]}
+
+
+def run(events, arm="D1", applicability="expected", plan=True):
+    with tempfile.TemporaryDirectory() as tmp:
+        root = Path(tmp); trajectory = root / "trajectory.jsonl"; trajectory.write_text("\n".join(json.dumps(e) for e in events)+"\n")
+        config = root / "artifacts/.deltawire/config.json"; config.parent.mkdir(parents=True, exist_ok=True); config.write_text("{}\n")
+        config_sha = __import__("hashlib").sha256(config.read_bytes()).hexdigest()
+        if plan:
+            p = root / "artifacts/.deltawire/plans/range-large.dw.json"; p.parent.mkdir(parents=True); p.write_text("{}\n")
+        expanded = [dict(item) for item in events]
+        expanded.insert(4, event("run_shell_command", {"command": "sha256sum .deltawire/config.json"}, config_sha))
+        expanded.insert(4, event("run_shell_command", {"command": "sha256sum /task-contract.json"}, CONTRACT_SHA))
+        trajectory.write_text("\n".join(json.dumps(e) for e in expanded)+"\n")
+        return receipt(trajectory, root, SHA, 1, arm, applicability, SCHEMA_SHA, config_sha, CONTRACT_SHA)
+
+
+GOOD = [
+    event("run_shell_command", {"command": "gemini skills list"}, "deltawire"), event("activate_skill", {"name": "deltawire"}),
+    event("run_shell_command", {"command": "command -v deltawire && deltawire version && sha256sum /usr/local/bin/deltawire"}, f"/usr/local/bin/deltawire\ndeltawire version dev\n{SHA}"),
+    event("read_file", {"file_path": "/task-contract.json"}, "{}"),
+    event("run_shell_command", {"command": "sha256sum .deltawire/schemas/range-large.schema.json"}, SCHEMA_SHA),
+    event("run_shell_command", {"command": "deltawire validate --repo . .deltawire/plans/range-large.dw.json"}),
+    event("run_shell_command", {"command": "verify_plan_contract.py --contract /task-contract.json --plan .deltawire/plans/range-large.dw.json"}),
+    event("run_shell_command", {"command": "deltawire render --repo . .deltawire/plans/range-large.dw.json"}),
+    event("run_shell_command", {"command": "deltawire check --repo . .deltawire/plans/range-large.dw.json"}),
+]
+
+
+def main():
+    assert run(GOOD)["classification"] == "used_successfully"
+    without_list = run(GOOD[1:])
+    assert without_list["classification"] != "used_successfully" and not without_list["skills_list_discovered"]
+    for index in range(1, len(GOOD)):
+        assert run(GOOD[:index] + GOOD[index+1:])["classification"] != "used_successfully"
+    with tempfile.TemporaryDirectory() as tmp:
+        root=Path(tmp); trajectory=root/"trajectory.jsonl"; trajectory.write_text(json.dumps(event("write_file", {}, "I used deltawire validate render check"))+"\n")
+        assert receipt(trajectory,root,SHA,1,"D1","expected",SCHEMA_SHA,"d"*64,CONTRACT_SHA)["classification"] == "not_used"
+    assert run(GOOD, arm="B0")["classification"] == "unexpected_use"
+    assert run([event("run_shell_command", {"command": "deltawire version"})], arm="B0", plan=False)["classification"] == "unexpected_use"
+    assert run([], arm="B0", plan=True)["classification"] == "unexpected_use"
+    abstain = [event("run_shell_command", {"command": "printf reason"}, "Too irregular for a compact deterministic plan")]
+    assert run(abstain, applicability="should_abstain", plan=False)["classification"] == "abstained"
+    with tempfile.TemporaryDirectory() as tmp:
+        root=Path(tmp); trajectory=root/"trajectory.jsonl"; trajectory.write_text("")
+        assert receipt(trajectory,root,SHA,1,"D1","should_abstain",SCHEMA_SHA,"d"*64,CONTRACT_SHA)["classification"] == "not_used"
+    failed = run(GOOD); failed["semantic_status"] = "fail"
+    with tempfile.TemporaryDirectory() as tmp:
+        root=Path(tmp); trajectory=root/"trajectory.jsonl"; p=root/"artifacts/.deltawire/plans/range.dw.json"; p.parent.mkdir(parents=True); p.write_text("{}\n")
+        config=root/"artifacts/.deltawire/config.json"; config.write_text("{}\n"); config_sha=__import__("hashlib").sha256(config.read_bytes()).hexdigest()
+        events=GOOD[:4]+[event("run_shell_command", {"command":"sha256sum /task-contract.json"}, CONTRACT_SHA),event("run_shell_command", {"command":"sha256sum .deltawire/config.json"}, config_sha)]+GOOD[4:]
+        trajectory.write_text("\n".join(json.dumps(e) for e in events)+"\n")
+        failed=receipt(trajectory,root,SHA,0,"D1","expected",SCHEMA_SHA,config_sha,CONTRACT_SHA)
+        assert failed["delivery_status"]=="pass" and failed["semantic_status"]=="fail" and failed["end_to_end_status"]=="fail"
+    print("Structured treatment-fidelity parser tests passed.")
+
+
+if __name__ == "__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/test_plan_contract.py b/labs/20-deltawire/eval/scripts/test_plan_contract.py
new file mode 100644
index 000000000..e60462c5b
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/test_plan_contract.py
@@ -0,0 +1,16 @@
+#!/usr/bin/env python3
+import json, shutil, subprocess, tempfile
+from pathlib import Path
+ROOT=Path(__file__).resolve().parents[4]; EVAL=ROOT/"labs/20-deltawire/eval"; CHECKER=EVAL/"scripts/verify_plan_contract.py"
+def plan(count=500,path="testdata/generated/output.ndjson",fmt="ndjson",schema=".deltawire/schemas/range-large.schema.json"):
+ return {"version":"deltawire.plan.v1","id":"range-large","record_schema":schema,"output":{"path":path,"format":fmt},"generators":[{"kind":"matrix","name":"range","dimensions":[{"name":"index","range":{"start":1,"end_exclusive":count+1,"step":1}}],"record":{"index":"${index}"}}],"assertions":{"count":count}}
+def main():
+ with tempfile.TemporaryDirectory() as tmp:
+  root=Path(tmp); binary=root/"deltawire"; subprocess.run(["go","build","-o",str(binary),"./cmd/deltawire"],cwd=ROOT/"labs/20-deltawire",check=True)
+  source=EVAL/"tasks/range-large"; shutil.copytree(source/"environment/.deltawire",root/".deltawire"); shutil.copyfile(source/"task-spec.json",root/"task-contract.json"); target=root/".deltawire/plans/range-large.dw.json"; target.parent.mkdir(parents=True)
+  for name,value,ok in [("correct",plan(),True),("count",plan(count=499),False),("path",plan(path="wrong.ndjson"),False),("format",plan(fmt="json"),False),("schema",plan(schema=".deltawire/schemas/wrong.json"),False)]:
+   target.write_text(json.dumps(value)+"\n"); receipt=root/f"{name}.json"; run=subprocess.run(["python3",str(CHECKER),"--contract",str(root/"task-contract.json"),"--plan",str(target),"--repo",str(root),"--deltawire",str(binary),"--receipt",str(receipt)],stdout=subprocess.DEVNULL)
+   assert (run.returncode==0)==ok and json.loads(receipt.read_text())["status"]==("pass" if ok else "fail")
+  target.write_text(json.dumps(plan())+"\n"); schema=root/".deltawire/schemas/range-large.schema.json"; schema.write_text(schema.read_text()+" "); run=subprocess.run(["python3",str(CHECKER),"--contract",str(root/"task-contract.json"),"--plan",str(target),"--repo",str(root),"--deltawire",str(binary),"--receipt",str(root/"mutated.json")],stdout=subprocess.DEVNULL); assert run.returncode
+ print("Plan-contract acceptance and rejection tests passed.")
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/test_runner_gates.py b/labs/20-deltawire/eval/scripts/test_runner_gates.py
new file mode 100644
index 000000000..6ca65ba58
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/test_runner_gates.py
@@ -0,0 +1,69 @@
+#!/usr/bin/env python3
+import hashlib
+import json
+import tempfile
+from pathlib import Path
+from unittest.mock import patch
+
+import runner
+
+
+def release_id(value):
+    body = dict(value)
+    body.pop("release_id", None)
+    return hashlib.sha256(json.dumps(body, sort_keys=True, separators=(",", ":")).encode()).hexdigest()
+
+
+def expect_runtime(fragment, function, *args):
+    try:
+        function(*args)
+        raise AssertionError(f"accepted invalid state: {fragment}")
+    except RuntimeError as error:
+        assert fragment in str(error), str(error)
+
+
+def main():
+    base = {
+        "schema_version": "deltawire-preflight-v3-manifest.v1",
+        "scope": ["probe/range-large-v3", "range-large/r1"],
+        "harbor_version": "0.20.0", "harbor_command": ["harbor"],
+        "model": "google/gemini-3.1-pro-preview", "agent": "gemini-cli", "input_hashes": {},
+        "deltawire_version": "deltawire version dev",
+    }
+    with tempfile.TemporaryDirectory() as tmp:
+        root = Path(tmp)
+        empty = root / "empty"
+        runner.refuse_nonempty(empty, "v3 probe")
+        for marker in ("partial", "failed", "successful"):
+            target = root / marker; target.mkdir(); (target / "evidence").write_text(marker)
+            expect_runtime("reruns are forbidden", runner.refuse_nonempty, target, "v3 probe")
+        path = root / "manifest.json"; path.write_text(json.dumps(base))
+        expect_runtime("only the dedicated v3 manifest", runner.load_approved_manifest, path, "0" * 64)
+    expect_runtime("frozen Harbor version mismatch", runner.verify_manifest, {**base, "harbor_version": "0.19.0"})
+    with patch("runner.subprocess.check_output", return_value="0.19.0\n"), patch("runner.verify_hash_lock"):
+        expect_runtime("Harbor version mismatch", runner.verify_manifest, base)
+
+    runs = runner.frozen_pair()
+    assert [run["arm"] for run in runs] == ["D1", "B0"]
+
+    with tempfile.TemporaryDirectory() as tmp:
+        root = Path(tmp); v3 = root / "results"; v3.mkdir(parents=True)
+        manifest = root / "manifest.json"; manifest.write_text("manifest\n")
+        with patch.object(runner, "V3_ROOT", v3):
+            expect_runtime("persisted probe-release", runner.verify_probe_release, manifest)
+            (v3 / "probe-release.json").write_text(json.dumps({"schema_version": "probe-release.v1", "status": "pass", "release_id": "fake"}))
+            expect_runtime("invalid or fabricated", runner.verify_probe_release, manifest)
+            release = {"schema_version": "probe-release.v1", "status": "pass", "manifest_sha256": runner.sha256(manifest),
+                       "evidence_hashes": {"missing.raw": "0" * 64}, "ready_for_72": False}
+            release["release_id"] = release_id(release)
+            (v3 / "probe-release.json").write_text(json.dumps(release))
+            with patch.object(runner, "ROOT", root):
+                expect_runtime("raw evidence mismatch", runner.verify_probe_release, manifest)
+
+    with patch("sys.argv", ["runner.py", "full", "--manifest", "x", "--approve-manifest-sha", "0" * 64, "--allow-72"]):
+        expect_runtime("unconditionally unauthorized", runner.main)
+    print("Versioned probe, immutable rerun, raw-release, adjacency, and full-mode gates passed.")
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/scripts/test_task_contracts.py b/labs/20-deltawire/eval/scripts/test_task_contracts.py
new file mode 100755
index 000000000..ecc6490ba
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/test_task_contracts.py
@@ -0,0 +1,61 @@
+#!/usr/bin/env python3
+import importlib.util
+import json
+import subprocess
+import tempfile
+import re
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[4]
+EVAL = ROOT / "labs/20-deltawire/eval"
+ORACLE_PATH = EVAL / "scripts/semantic_oracle.py"
+module_spec = importlib.util.spec_from_file_location("semantic_oracle", ORACLE_PATH)
+oracle = importlib.util.module_from_spec(module_spec); module_spec.loader.exec_module(oracle)
+CASES = ("canonical", "empty", "malformed", "missing", "duplicate", "extra", "wrong_value", "wrong_field_set", "family_mut")
+
+
+def write_ndjson(path, rows):
+    path.write_text("".join(json.dumps(row) + "\n" for row in rows), encoding="utf-8")
+
+
+def mutate(rows, case):
+    result = [dict(row) for row in rows]
+    if case == "missing": return result[:-1]
+    if case == "duplicate": return result + [dict(result[0])]
+    if case == "extra": return result + [{"unexpected": True}]
+    if case == "wrong_field_set": result[0]["unexpected"] = True; return result
+    if case == "family_mut": return list(reversed(result))
+    if case == "wrong_value":
+        key = next(iter(result[0]))
+        value = result[0][key]
+        result[0][key] = (not value) if isinstance(value, bool) else (value + 999999 if isinstance(value, int) else "__wrong__")
+        return result
+    return result
+
+
+def result(spec_path, output_path):
+    run = subprocess.run(["python3", str(ORACLE_PATH), str(spec_path), str(output_path)], capture_output=True, text=True, check=True)
+    return json.loads(run.stdout)["exact_match"]
+
+
+def main():
+    with tempfile.TemporaryDirectory(prefix="deltawire-contracts-") as tmp:
+        temp = Path(tmp)
+        for spec_path in sorted((EVAL / "tasks").glob("*/task-spec.json")):
+            spec = json.loads(spec_path.read_text(encoding="utf-8")); rows = oracle.canonical(spec)
+            task=spec_path.parent; match=re.search(r"```json\n(.*?)```",(task/"instruction.md").read_text(),re.DOTALL)
+            assert spec==json.loads((task/"tests/task-spec.json").read_text())==json.loads((task/"environment/task-contract.json").read_text())==json.loads(match.group(1))
+            assert "tests/" not in (task/"instruction.md").read_text()
+            for case in CASES:
+                output = temp / f"{spec['task']}-{case}.ndjson"
+                if case == "empty": output.write_text("", encoding="utf-8")
+                elif case == "malformed": output.write_text("{not-json}\n", encoding="utf-8")
+                else: write_ndjson(output, rows if case == "canonical" else mutate(rows, case))
+                observed = result(spec_path, output)
+                expected = 1 if case == "canonical" else 0
+                if observed != expected:
+                    raise SystemExit(f"{spec['task']} {case}: expected {expected}, got {observed}")
+    print("All projections match; 12 canonical contracts and 96 mutation cases passed.")
+
+
+if __name__ == "__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/v4/build_and_stage.sh b/labs/20-deltawire/eval/scripts/v4/build_and_stage.sh
new file mode 100644
index 000000000..98f46d802
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/build_and_stage.sh
@@ -0,0 +1,11 @@
+#!/usr/bin/env bash
+set -euo pipefail
+ROOT="$(cd "$(dirname "$0")/../../../../.." && pwd)"
+LAB="$ROOT/labs/20-deltawire"; GENERATED="$LAB/eval/.generated"; mkdir -p "$GENERATED"
+(cd "$LAB" && CGO_ENABLED=0 GOOS=linux GOARCH=amd64 GOWORK=off go build -trimpath -buildvcs=false -o "$GENERATED/deltawire-linux-amd64" ./cmd/deltawire)
+SHA="$(sha256sum "$GENERATED/deltawire-linux-amd64" | cut -d' ' -f1)"
+for task in "$LAB/eval/probes/range-large-v4" "$LAB/eval/probes/range-large-v4/pair-task"; do
+ mkdir -p "$task/environment/.generated"; cp "$GENERATED/deltawire-linux-amd64" "$task/environment/.generated/deltawire"
+ test "$(sha256sum "$task/environment/.generated/deltawire" | cut -d' ' -f1)" = "$SHA"
+done
+printf '%s\n' "$SHA"
diff --git a/labs/20-deltawire/eval/scripts/v4/environment_receipt.py b/labs/20-deltawire/eval/scripts/v4/environment_receipt.py
new file mode 100644
index 000000000..b973414ba
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/environment_receipt.py
@@ -0,0 +1,36 @@
+#!/usr/bin/env python3
+"""Generate deterministic DeltaWire infrastructure evidence inside the task."""
+import argparse,hashlib,json,os,stat,subprocess
+from pathlib import Path
+
+def sha(path):
+ h=hashlib.sha256()
+ with Path(path).open("rb") as f:
+  for chunk in iter(lambda:f.read(1048576),b""): h.update(chunk)
+ return h.hexdigest()
+def main():
+ p=argparse.ArgumentParser(); p.add_argument("--expectations",required=True); p.add_argument("--receipt",required=True); a=p.parse_args()
+ expected=json.loads(Path(a.expectations).read_text()); binary=Path(expected["binary"]["path"])
+ exists=binary.exists(); lst=binary.lstat() if exists else None
+ is_regular=bool(lst and stat.S_ISREG(lst.st_mode)); is_symlink=binary.is_symlink() if exists else False
+ executable=exists and os.access(binary,os.X_OK); realpath=str(binary.resolve()) if exists else None
+ try: version=subprocess.run([str(binary),"version"],capture_output=True,text=True,check=False) if exists else None
+ except OSError: version=None
+ files={}
+ for name,item in expected["files"].items():
+  path=Path(item["path"]); actual=sha(path) if path.is_file() else None
+  files[name]={"path":item["path"],"exists":path.is_file(),"actual_sha256":actual,"expected_sha256":item["sha256"],"hash_match":actual==item["sha256"]}
+ binary_hash=sha(binary) if is_regular else None
+ checks={
+  "binary_exists":exists,"binary_regular":is_regular,"binary_not_symlink":not is_symlink,
+  "binary_executable":executable,"binary_path":str(binary)==expected["binary"]["path"],
+  "binary_realpath":realpath==expected["binary"]["realpath"],
+  "binary_hash":binary_hash==expected["binary"]["sha256"],
+  "version_exit_0":bool(version and version.returncode==0),
+  "version_exact":bool(version and version.stdout.splitlines() and version.stdout.splitlines()[0].strip()==expected["binary"]["version"]),
+  **{f"{name}_hash":item["hash_match"] for name,item in files.items()},
+ }
+ receipt={"schema_version":"deltawire-environment-receipt.v1","binary":{"path":str(binary),"realpath":realpath,"exists":exists,"is_regular":is_regular,"is_symlink":is_symlink,"executable":executable,"actual_sha256":binary_hash,"expected_sha256":expected["binary"]["sha256"],"version_stdout":version.stdout.strip() if version else None,"version_stderr":version.stderr.strip() if version else None,"version_exit_code":version.returncode if version else None,"expected_version":expected["binary"]["version"]},"files":files,"checks":checks,"status":"pass" if all(checks.values()) else "fail"}
+ target=Path(a.receipt); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(receipt,indent=2,sort_keys=True)+"\n")
+ print(json.dumps(receipt,sort_keys=True)); raise SystemExit(0 if receipt["status"]=="pass" else 1)
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/v4/fidelity.py b/labs/20-deltawire/eval/scripts/v4/fidelity.py
new file mode 100644
index 000000000..6d72cac6f
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/fidelity.py
@@ -0,0 +1,36 @@
+#!/usr/bin/env python3
+import hashlib,json
+from pathlib import Path
+from shell_observation import calls,parse,invokes_deltawire
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def ok_tool(call): return call.get("tool_status")=="success"
+def build(trajectory,trial,arm,spec,semantic,independent_receipt):
+ trial=Path(trial); artifacts=trial/"artifacts"; structured=calls(trajectory); observations,lifecycle=parse(trajectory)
+ obs_by_command={o["command"]:o for o in observations}
+ discovery=any(c["name"] in {"run_shell_command","shell","execute"} and "gemini skills list" in str(c["args"].get("command") or "") and "deltawire" in c["output"].lower() and ok_tool(c) and obs_by_command.get(str(c["args"].get("command") or ""),{}).get("shell_exit_code") in (None,0) for c in structured)
+ activation=any(c["name"]=="activate_skill" and str(c["args"].get("name") or c["args"].get("skill_name") or "").lower()=="deltawire" and ok_tool(c) for c in structured)
+ contract_access=any(ok_tool(c) and ((c["name"] in {"read_file","read_many_files"} and "/task-contract.json" in json.dumps(c["args"])) or (c["name"] in {"run_shell_command","shell","execute"} and "/task-contract.json" in str(c["args"].get("command") or ""))) for c in structured)
+ by_name={name:[x for x in lifecycle if x["command"]==name] for name in ("validate","render","check")}
+ lifecycle_observed={name:bool(items) and any(item["shell_exit_code"]!=0 or item["shell_exit_code"]==0 for item in items) for name,items in by_name.items()}
+ # An explicit nonzero final attempt fails; null is resolved by retained independent evidence.
+ lifecycle_not_proven_failed={name:(items[-1]["shell_exit_code"]==0 if items and items[-1]["shell_exit_code"] is not None else bool(items)) for name,items in by_name.items()}
+ paths={
+  "plan":artifacts/".deltawire/plans/range-large.dw.json","config":artifacts/".deltawire/config.json",
+  "schema":artifacts/".deltawire/schemas/range-large.schema.json","state":artifacts/".deltawire/state.json",
+  "output":artifacts/"testdata/generated/output.ndjson","in_container_contract":artifacts/".deltawire/plan-contract-receipt.json",
+ }
+ retained={name:path.is_file() for name,path in paths.items()}
+ in_container=retained["in_container_contract"] and json.loads(paths["in_container_contract"].read_text()).get("status")=="pass"
+ independent=Path(independent_receipt).is_file() and json.loads(Path(independent_receipt).read_text()).get("status")=="pass"
+ schema_integrity=retained["schema"] and sha(paths["schema"])==spec["authoritative_schema"]["sha256"]
+ deltawire_invocations=[obs for obs in observations if invokes_deltawire(obs)]
+ contamination=arm=="B0" and (activation or bool(deltawire_invocations) or retained["plan"] or retained["state"])
+ checks={"structured_discovery":discovery,"structured_activation":activation,"public_contract_access":contract_access,
+         **{f"{name}_invoked":lifecycle_observed[name] for name in lifecycle_observed},
+         **{f"{name}_not_proven_failed":lifecycle_not_proven_failed[name] for name in lifecycle_not_proven_failed},
+         **{f"retained_{name}":value for name,value in retained.items()},"in_container_contract_pass":in_container,
+         "independent_contract_pass":independent,"retained_schema_integrity":schema_integrity,"semantic_exact_pass":semantic==1}
+ if arm=="D1": status="pass" if all(checks.values()) else "fail"; classification="used_successfully" if status=="pass" else "attempted_failed"
+ else: status="pass" if not contamination and semantic==1 else "fail"; classification="not_used" if not contamination else "contaminated"
+ return {"schema_version":"treatment-use-receipt.v4","arm":arm,"classification":classification,"status":status,"checks":checks,"shell_observations":observations,"lifecycle_invocations":lifecycle,"contamination":contamination,"artifact_hashes":{name:sha(path) for name,path in paths.items() if path.is_file()},"trajectory_sha256":sha(trajectory)}
diff --git a/labs/20-deltawire/eval/scripts/v4/generate_manifest.py b/labs/20-deltawire/eval/scripts/v4/generate_manifest.py
new file mode 100644
index 000000000..2bb2e8b2e
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/generate_manifest.py
@@ -0,0 +1,15 @@
+#!/usr/bin/env python3
+import argparse,hashlib,json
+from pathlib import Path
+ROOT=Path(__file__).resolve().parents[5];EVAL=ROOT/"labs/20-deltawire/eval";OUT=EVAL/"manifests/preflight-v4-range-large.json"
+def sha(p):return hashlib.sha256(p.read_bytes()).hexdigest()
+def main():
+ p=argparse.ArgumentParser();p.add_argument("--input-commit",required=True);p.add_argument("--binary-sha",required=True);a=p.parse_args()
+ paths=[EVAL/"matrices/72-run-matrix.json",EVAL/"tasks/range-large/task-spec.json",EVAL/"manifests/preflight-v2-evidence-lock.json",EVAL/"manifests/preflight-v3-evidence-lock.json",EVAL/"results/preflight-v1/treatment-probe-range-large/v1-evidence-lock.json"]
+ for base in (EVAL/"scripts/v4",EVAL/"probes/range-large-v4"):
+  paths += [x for x in base.rglob("*") if x.is_file() and ".generated" not in x.parts and "__pycache__" not in x.parts and x.suffix!=".pyc" and x.name!="preflight-v4-range-large.json"]
+ expectation=json.loads((EVAL/"probes/range-large-v4/environment/deltawire-environment-expectations.json").read_text())
+ if expectation["binary"]["sha256"]!=a.binary_sha:raise SystemExit("environment expectation binary hash mismatch")
+ value={"schema_version":"deltawire-preflight-v4-manifest.v1","repository_commit":a.input_commit,"scope":["probe/range-large-v4","range-large/r1"],"matrix_positions":{"zero_based":[36,37],"order":["D1","B0"]},"deltawire_binary_sha256":a.binary_sha,"model":"google/gemini-3.1-pro-preview","expected_reported_model":"gemini-3.1-pro-preview","agent":"gemini-cli","harbor_version":"0.20.0","harbor_package":"harbor==0.20.0","harbor_command":["uvx","--from","harbor==0.20.0","harbor"],"input_hashes":{str(x.relative_to(ROOT)):sha(x) for x in sorted(set(paths))},"failure_policy":"One v4 probe; one automatically released range-large/r1 pair only after probe-release.v2; never rerun or run full.","ready_for_72":False}
+ OUT.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n");print(OUT.relative_to(ROOT))
+if __name__=="__main__":main()
diff --git a/labs/20-deltawire/eval/scripts/v4/generate_v3_evidence_lock.py b/labs/20-deltawire/eval/scripts/v4/generate_v3_evidence_lock.py
new file mode 100644
index 000000000..c8093c14b
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/generate_v3_evidence_lock.py
@@ -0,0 +1,39 @@
+#!/usr/bin/env python3
+"""Lock immutable v3 evidence and source bytes from the evidence commit."""
+import hashlib, json, subprocess
+from pathlib import Path
+
+ROOT=Path(__file__).resolve().parents[5]
+EVAL=ROOT/"labs/20-deltawire/eval"
+COMMIT="07e341fd213b6deb17cce9115263cebe8ba2105f"
+OUT=EVAL/"manifests/preflight-v3-evidence-lock.json"
+
+def git(*args): return subprocess.check_output(["git",*args],cwd=ROOT)
+def content(path): return git("show",f"{COMMIT}:{path}")
+def tree(prefix):
+    raw=git("ls-tree","-r","--name-only",COMMIT,"--",prefix).decode()
+    return {line for line in raw.splitlines() if line}
+
+def main():
+    manifest_path="labs/20-deltawire/eval/manifests/preflight-v3-range-large.json"
+    manifest=json.loads(content(manifest_path))
+    paths=set(manifest["input_hashes"])|{manifest_path}
+    paths|=tree("labs/20-deltawire/eval/probes/range-large-v3")
+    paths|=tree("labs/20-deltawire/eval/results/preflight-v3")
+    paths|={
+      "labs/20-deltawire/eval/scripts/runner.py",
+      "labs/20-deltawire/eval/scripts/fidelity.py",
+      "labs/20-deltawire/eval/scripts/test_fidelity.py",
+      "labs/20-deltawire/eval/scripts/test_runner_gates.py",
+      "labs/20-deltawire/eval/scripts/validate_preflight.py",
+      "labs/20-deltawire/eval/scripts/generate_preflight_v3_manifest.py",
+    }
+    value={
+      "schema_version":"git-evidence-lock.v1",
+      "source_commit":COMMIT,
+      "source_tree":git("rev-parse",f"{COMMIT}^{{tree}}").decode().strip(),
+      "immutable_files":{p:hashlib.sha256(content(p)).hexdigest() for p in sorted(paths)},
+    }
+    OUT.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+    print(OUT.relative_to(ROOT),len(paths))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/v4/runner.py b/labs/20-deltawire/eval/scripts/v4/runner.py
new file mode 100644
index 000000000..32583153d
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/runner.py
@@ -0,0 +1,162 @@
+#!/usr/bin/env python3
+"""One-shot DeltaWire v4 measurement-recovery runner."""
+import argparse,hashlib,json,os,shlex,subprocess,sys,tempfile
+from datetime import datetime
+from pathlib import Path
+from fidelity import build,sha
+
+ROOT=Path(__file__).resolve().parents[5]; EVAL=ROOT/"labs/20-deltawire/eval"; HERE=Path(__file__).resolve().parent
+PROBE=EVAL/"probes/range-large-v4"; PAIR_TASK=PROBE/"pair-task"; SKILL=PROBE/"skill/deltawire"
+RESULTS=EVAL/"results/preflight-v4"; PROBE_ROOT=RESULTS/"treatment-probe-range-large-v4"; PAIR_ROOT=RESULTS/"range-large-r1"
+MANIFEST=EVAL/"manifests/preflight-v4-range-large.json"; BINARY=EVAL/".generated/deltawire-linux-amd64"
+
+def load(p):return json.loads(Path(p).read_text(encoding="utf-8"))
+def write(p,v):p=Path(p);p.parent.mkdir(parents=True,exist_ok=True);p.write_text(json.dumps(v,indent=2,sort_keys=True)+"\n")
+def rel(p):return str(Path(p).resolve().relative_to(ROOT.resolve()))
+def refuse(root,label):
+ if root.exists() and any(root.iterdir()):raise RuntimeError(f"{label} result root is nonempty; rerun forbidden")
+def verify_locks():subprocess.run(["python3",str(HERE/"verify_evidence_locks.py")],cwd=ROOT,check=True)
+def verify_manifest(path,approved):
+ path=Path(path).resolve()
+ if path!=MANIFEST.resolve() or sha(path)!=approved:raise RuntimeError("v4 manifest path or SHA mismatch")
+ m=load(path)
+ if m.get("schema_version")!="deltawire-preflight-v4-manifest.v1" or m.get("scope")!=["probe/range-large-v4","range-large/r1"]:raise RuntimeError("v4 manifest identity mismatch")
+ if m.get("harbor_version")!="0.20.0" or m.get("model")!="google/gemini-3.1-pro-preview" or m.get("agent")!="gemini-cli":raise RuntimeError("frozen runtime mismatch")
+ actual=subprocess.check_output(m["harbor_command"]+["--version"],cwd=ROOT,text=True).strip()
+ if actual!="0.20.0":raise RuntimeError("Harbor version mismatch")
+ for name,want in m["input_hashes"].items():
+  target=ROOT/name
+  if not target.is_file() or sha(target)!=want:raise RuntimeError(f"frozen input mismatch: {name}")
+ verify_locks();return m
+def verify_binary(paths,want):
+ for path in [BINARY,*[p/"environment/.generated/deltawire" for p in paths]]:
+  if not path.is_file() or sha(path)!=want:raise RuntimeError(f"binary parity mismatch: {path}")
+ version=subprocess.run([str(BINARY),"version"],capture_output=True,text=True,check=False)
+ if version.returncode or not version.stdout.startswith("deltawire version dev\n"):raise RuntimeError("host binary version mismatch")
+
+def private_env():
+ key=os.environ.get("GEMINI_API_KEY","").strip()
+ if not key:raise RuntimeError("GEMINI_API_KEY is required")
+ f=tempfile.NamedTemporaryFile("w",prefix="deltawire-v4-",suffix=".env",delete=False);f.write("GEMINI_API_KEY="+key.replace("\n","")+"\n");f.close();os.chmod(f.name,0o600);return Path(f.name)
+def trials(jobs):return {p.resolve() for p in jobs.glob("*/*__*") if p.is_dir()}
+def run_one(m,run,task,jobs,dry=False):
+ cmd=m["harbor_command"]+["run","-a",m["agent"],"-m",m["model"],"-p",str(task),"-k","1","--jobs-dir",str(jobs),"--n-concurrent","1"]
+ if run["arm"]=="D1":cmd+=["--skill",str(SKILL)]
+ redacted=shlex.join(cmd+["--env-file","<private-env-file>"])
+ if dry:print(redacted);return {**run,"status":"dry_run","redacted_command":redacted}
+ envfile=private_env();before=trials(jobs);jobs.mkdir(parents=True,exist_ok=True);env=os.environ.copy();env.pop("GEMINI_API_KEY",None)
+ try:
+  try:rc=subprocess.run(cmd+["--env-file",str(envfile)],cwd=ROOT,env=env,check=False).returncode;error=None
+  except OSError as e:rc=None;error=str(e)
+ finally:envfile.unlink(missing_ok=True)
+ created=trials(jobs)-before;entry={**run,"harbor_exit_code":rc,"status":"completed" if rc==0 else "harbor_failed","redacted_command":redacted}
+ if error:entry["harbor_error"]=error
+ if len(created)!=1:entry["artifact_error"]=f"expected one trial, found {len(created)}";return entry
+ trial=created.pop();entry["trial_dir"]=rel(trial)
+ candidates={"result_path":trial/"result.json","trajectory_path":trial/"agent/gemini-cli.trajectory.jsonl"}
+ outputs=sorted((trial/"artifacts").rglob("output.ndjson"))
+ if outputs:candidates["output_path"]=outputs[0]
+ for key,path in candidates.items():
+  if path.is_file():entry[key]=rel(path);entry[key.replace("_path","_sha256")]=sha(path)
+ missing=[k for k in candidates if k not in entry]
+ if missing:entry["artifact_error"]="missing artifacts: "+", ".join(missing)
+ return entry
+
+def semantic(output):
+ run=subprocess.run(["python3",str(HERE/"semantic_oracle.py"),str(EVAL/"tasks/range-large/task-spec.json"),str(output)],capture_output=True,text=True,check=False)
+ try:value=json.loads(run.stdout) if run.stdout.strip() else {}
+ except json.JSONDecodeError:value={}
+ exact=value.get("exact_match",0) if run.returncode==0 else 0
+ return {"schema_version":"semantic-result.v4","exact_match":exact,"status":"pass" if exact==1 else "fail","oracle_result":value,"error":None if exact==1 else (run.stderr.strip() or run.stdout.strip() or "oracle failed")}
+def independent(trial,dest):
+ plans=sorted((Path(trial)/"artifacts/.deltawire/plans").glob("*.dw.json"))
+ if len(plans)!=1:value={"schema_version":"plan-contract-receipt.v1","status":"fail","checks":{},"error":f"expected one plan, found {len(plans)}"};write(dest,value);return value
+ cmd=["python3",str(HERE/"verify_plan_contract.py"),"--contract",str(EVAL/"tasks/range-large/task-spec.json"),"--plan",str(plans[0]),"--repo",str(Path(trial)/"artifacts"),"--deltawire",str(BINARY),"--receipt",str(dest)]
+ subprocess.run(cmd,cwd=ROOT,capture_output=True,text=True,check=False)
+ return load(dest) if Path(dest).is_file() else {"status":"fail"}
+def duration(data):
+ try:return (datetime.fromisoformat(data["finished_at"].replace("Z","+00:00"))-datetime.fromisoformat(data["started_at"].replace("Z","+00:00"))).total_seconds()
+ except (KeyError,TypeError,ValueError):return None
+def red_receipt(arm,reason):
+ return {"schema_version":"treatment-use-receipt.v4","arm":arm,"status":"fail","classification":"attempted_failed" if arm=="D1" else "contaminated","failure_reasons":[reason],"checks":{},"contamination":arm=="B0"}
+def finalize(m,entry,root):
+ arm=entry["arm"];trial=ROOT/entry["trial_dir"] if entry.get("trial_dir") else None;trajectory=ROOT/entry["trajectory_path"] if entry.get("trajectory_path") else None;output=ROOT/entry["output_path"] if entry.get("output_path") else None
+ sem=semantic(output) if output and output.is_file() else {"schema_version":"semantic-result.v4","exact_match":0,"status":"fail","error":"output missing"};sem_path=root/f"{arm}-semantic-result.json";write(sem_path,sem)
+ indep_path=root/f"{arm}-plan-contract-receipt.json"
+ if arm=="D1" and trial:indep=independent(trial,indep_path)
+ else:indep={"status":"not_applicable"}
+ spec=load(EVAL/"tasks/range-large/task-spec.json")
+ try:use=build(trajectory,trial,arm,spec,sem["exact_match"],indep_path) if trajectory and trial else red_receipt(arm,entry.get("artifact_error","trial missing"))
+ except Exception as e:use=red_receipt(arm,str(e))
+ use_path=root/f"{arm}-treatment-use.json";write(use_path,use)
+ environment_path=trial/"artifacts/.deltawire/environment-receipt.json" if trial else None
+ environment=load(environment_path) if environment_path and environment_path.is_file() else {"schema_version":"deltawire-environment-receipt.v1","status":"fail","error":"environment receipt missing"}
+ environment_copy=root/f"{arm}-environment-receipt.json";write(environment_copy,environment)
+ end={"schema_version":"deltawire-end-to-end-receipt.v4","arm":arm,"environment_status":environment.get("status"),"treatment_status":use.get("status"),"semantic_status":sem.get("status"),"status":"pass" if environment.get("status")==use.get("status")==sem.get("status")=="pass" else "fail","ready_for_72":False,"source_hashes":{"environment":sha(environment_copy),"treatment_use":sha(use_path),"semantic":sha(sem_path),"trajectory":entry.get("trajectory_sha256"),"result":entry.get("result_sha256"),"output":entry.get("output_sha256")}}
+ data=load(ROOT/entry["result_path"]) if entry.get("result_path") else {};agent=data.get("agent_result") or {};entry.update({"trial_id":data.get("id") or data.get("trial_name"),"reported_model":((data.get("agent_info") or {}).get("model_info") or {}).get("name"),"input_tokens":agent.get("n_input_tokens"),"output_tokens":agent.get("n_output_tokens"),"cache_tokens":agent.get("n_cache_tokens"),"wall_time_seconds":duration(data)})
+ end["model_match"]=entry["reported_model"]==m["expected_reported_model"];end["harbor_exit_0"]=entry.get("harbor_exit_code")==0
+ if not end["model_match"] or not end["harbor_exit_0"] or entry.get("artifact_error"):end["status"]="fail"
+ end_path=root/f"{arm}-end-to-end.json";write(end_path,end);return sem,use,environment,end,{"semantic":sem_path,"use":use_path,"environment":environment_copy,"end":end_path,"independent":indep_path}
+def finalize_failure(entry,root,error):
+ arm=entry["arm"];reason=str(error);sem={"schema_version":"semantic-result.v4","exact_match":0,"status":"fail","error":reason};use=red_receipt(arm,reason);env={"schema_version":"deltawire-environment-receipt.v1","status":"fail","error":reason}
+ sem_path=root/f"{arm}-semantic-result.json";use_path=root/f"{arm}-treatment-use.json";env_path=root/f"{arm}-environment-receipt.json";indep_path=root/f"{arm}-plan-contract-receipt.json"
+ write(sem_path,sem);write(use_path,use);write(env_path,env)
+ if arm=="D1":write(indep_path,{"schema_version":"plan-contract-receipt.v1","status":"fail","checks":{},"error":reason})
+ end={"schema_version":"deltawire-end-to-end-receipt.v4","arm":arm,"environment_status":"fail","treatment_status":"fail","semantic_status":"fail","status":"fail","ready_for_72":False,"failure_reasons":[reason],"source_hashes":{"environment":sha(env_path),"treatment_use":sha(use_path),"semantic":sha(sem_path),"trajectory":entry.get("trajectory_sha256"),"result":entry.get("result_sha256"),"output":entry.get("output_sha256")}}
+ write(root/f"{arm}-end-to-end.json",end);return end
+
+def release_body(manifest,ledger,entry,paths):
+ trial=ROOT/entry["trial_dir"];required=[Path(manifest),Path(ledger),ROOT/entry["result_path"],ROOT/entry["trajectory_path"],ROOT/entry["output_path"],trial/"artifacts/.deltawire/config.json",trial/"artifacts/.deltawire/schemas/range-large.schema.json",trial/"artifacts/.deltawire/plans/range-large.dw.json",trial/"artifacts/.deltawire/state.json",trial/"artifacts/.deltawire/plan-contract-receipt.json",*paths.values()]
+ missing=[str(p) for p in required if not p.is_file()]
+ if missing:raise RuntimeError("release artifacts missing: "+", ".join(missing))
+ body={"schema_version":"probe-release.v2","probe":"range-large-v4","status":"pass","manifest_sha256":sha(manifest),"evidence_hashes":{rel(p):sha(p) for p in required},"ready_for_72":False}
+ body["release_id"]=hashlib.sha256(json.dumps(body,sort_keys=True,separators=(",",":")).encode()).hexdigest();return body
+def verify_release(manifest):
+ path=PROBE_ROOT/"probe-release.json"
+ if not path.is_file():raise RuntimeError("probe-release.v2 missing")
+ release=load(path);rid=release.pop("release_id",None);want=hashlib.sha256(json.dumps(release,sort_keys=True,separators=(",",":")).encode()).hexdigest();release["release_id"]=rid
+ if release.get("schema_version")!="probe-release.v2" or release.get("status")!="pass" or rid!=want or release.get("manifest_sha256")!=sha(manifest):raise RuntimeError("invalid probe release")
+ for name,want_hash in release["evidence_hashes"].items():
+  target=ROOT/name
+  if not target.is_file() or sha(target)!=want_hash:raise RuntimeError(f"release evidence mismatch: {name}")
+ ledger=load(PROBE_ROOT/"run-ledger.json");entry=ledger["runs"][0]
+ if semantic(ROOT/entry["output_path"])["exact_match"]!=1:raise RuntimeError("release semantic recheck failed")
+ with tempfile.TemporaryDirectory() as tmp:
+  if independent(ROOT/entry["trial_dir"],Path(tmp)/"receipt.json").get("status")!="pass":raise RuntimeError("release contract recheck failed")
+ env=load(PROBE_ROOT/"D1-environment-receipt.json");use=load(PROBE_ROOT/"D1-treatment-use.json");end=load(PROBE_ROOT/"D1-end-to-end.json")
+ if not all(x.get("status")=="pass" for x in (env,use,end)):raise RuntimeError("release receipt status failed")
+ rerun=build(ROOT/entry["trajectory_path"],ROOT/entry["trial_dir"],"D1",load(EVAL/"tasks/range-large/task-spec.json"),1,PROBE_ROOT/"D1-plan-contract-receipt.json")
+ if rerun.get("status")!="pass":raise RuntimeError("release trajectory recheck failed")
+ return release
+def frozen_pair():
+ matrix=load(EVAL/"matrices/72-run-matrix.json");runs=[matrix[36],matrix[37]]
+ if [r.get("pair_id") for r in runs]!=["range-large/r1"]*2 or [r.get("arm") for r in runs]!=["D1","B0"]:raise RuntimeError("frozen matrix positions 36-37 mismatch")
+ return runs
+def args():
+ p=argparse.ArgumentParser();p.add_argument("mode",choices=("probe-v4","pair"));p.add_argument("--manifest",required=True);p.add_argument("--approve-manifest-sha",required=True);p.add_argument("--pair-id");p.add_argument("--dry-run",action="store_true");return p.parse_args()
+def main():
+ a=args();m=verify_manifest(a.manifest,a.approve_manifest_sha)
+ if a.mode=="probe-v4":runs=[{"pair_id":"probe/range-large-v4","task":"range-large-v4","arm":"D1","repetition":1}];tasks=[PROBE];root=PROBE_ROOT
+ else:
+  if a.pair_id!="range-large/r1":raise RuntimeError("only range-large/r1 is permitted")
+  verify_release(a.manifest);runs=frozen_pair();tasks=[PAIR_TASK,PAIR_TASK];root=PAIR_ROOT
+ if not a.dry_run:refuse(root,a.mode)
+ verify_binary(tasks,m["deltawire_binary_sha256"]);ledger={"schema_version":"run-ledger.v4","mode":a.mode,"ready_for_72":False,"runs":[]};ledger_path=root/"run-ledger.json"
+ for run,task in zip(runs,tasks):
+  entry=run_one(m,run,task,root/"raw"/run["arm"],a.dry_run);ledger["runs"].append(entry)
+  if a.dry_run:continue
+  write(ledger_path,ledger)
+  try:sem,use,env,end,paths=finalize(m,entry,root)
+  except Exception as error:end=finalize_failure(entry,root,error);write(ledger_path,ledger);raise RuntimeError(f"{a.mode} {run['arm']} finalization failed: {error}")
+  write(ledger_path,ledger)
+  if end["status"]!="pass":raise RuntimeError(f"{a.mode} {run['arm']} end-to-end gate failed")
+  if a.mode=="probe-v4":
+   release=release_body(a.manifest,ledger_path,entry,paths);write(root/"probe-release.json",release)
+   write(root/"pre-pair-report.json",{"schema_version":"pre-pair-report.v1","probe_status":"pass","release_id":release["release_id"],"manifest_sha256":sha(a.manifest),"ready_for_72":False,"automatic_pair_authorized":True})
+ if a.dry_run:return
+ if a.mode=="pair":
+  d1,b0=ledger["runs"];write(root/"pair-summary.json",{"schema_version":"operational-positive-control.v1","pair_complete":True,"order":["D1","B0"],"D1":{k:d1.get(k) for k in ("trial_id","input_tokens","output_tokens","cache_tokens","wall_time_seconds")},"B0":{k:b0.get(k) for k in ("trial_id","input_tokens","output_tokens","cache_tokens","wall_time_seconds")},"claim":"operational_positive_control_only","READY_FOR_72":False,"full_run_started":False})
+ print(a.mode,"completed")
+if __name__=="__main__":
+ try:main()
+ except (OSError,ValueError,KeyError,RuntimeError,subprocess.CalledProcessError) as e:print("ERROR:",e,file=sys.stderr);raise SystemExit(1)
diff --git a/labs/20-deltawire/eval/scripts/v4/semantic_oracle.py b/labs/20-deltawire/eval/scripts/v4/semantic_oracle.py
new file mode 100755
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/scripts/v4/shell_observation.py b/labs/20-deltawire/eval/scripts/v4/shell_observation.py
new file mode 100644
index 000000000..1d816d864
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/shell_observation.py
@@ -0,0 +1,70 @@
+#!/usr/bin/env python3
+"""Canonicalize structured Gemini shell calls without trusting tool status."""
+import json,re,shlex
+from pathlib import Path
+
+SHELL_NAMES={"run_shell_command","shell","execute"}
+EXIT_KEYS=("exit_code","exitCode","returncode","return_code")
+TRAILER=re.compile(r"(?:^|\n)Exit Code:\s*(-?\d+)\nProcess Group PGID:\s*\d+\s*(?:</untrusted_context>)?\s*$")
+
+def calls(path):
+ out=[]
+ for line in Path(path).read_text(encoding="utf-8",errors="replace").splitlines():
+  try: event=json.loads(line)
+  except json.JSONDecodeError: continue
+  for call in event.get("toolCalls",[]):
+   outputs=[]
+   for item in call.get("result",[]):
+    response=(item.get("functionResponse") or {}).get("response") or {}
+    outputs.append(str(response.get("output","")))
+   out.append({"name":call.get("name"),"args":call.get("args") or {},"tool_status":call.get("status"),"result":call.get("result") or [],"output":"\n".join(outputs)})
+ return out
+
+def direct_code(value):
+ if isinstance(value,dict):
+  for key in EXIT_KEYS:
+   if key in value and isinstance(value[key],int): return value[key]
+  for child in value.values():
+   found=direct_code(child)
+   if found is not None:return found
+ elif isinstance(value,list):
+  for child in value:
+   found=direct_code(child)
+   if found is not None:return found
+ return None
+
+def observation(call):
+ command=str(call["args"].get("command") or call["args"].get("cmd") or "")
+ code=direct_code(call["result"]); provenance="structured" if code is not None else None
+ if code is None:
+  matches=TRAILER.findall(call["output"])
+  if len(matches)==1: code=int(matches[0]); provenance="explicit_output"
+ return {"command":command,"tool_status":call["tool_status"],"shell_exit_code":code,"shell_exit_provenance":provenance,"observable":call["name"] in SHELL_NAMES and bool(command)}
+
+def lifecycle(observation):
+ command=observation["command"]; parts=re.split(r"(&&|\|\||;|\n)",command); commands=parts[::2]; operators=parts[1::2]
+ found=[]
+ for segment in commands:
+  try: tokens=shlex.split(segment)
+  except ValueError: continue
+  while tokens and ("=" in tokens[0] and not tokens[0].startswith("/")): tokens=tokens[1:]
+  if tokens and tokens[0] in {"sudo","env"}: tokens=tokens[1:]
+  if len(tokens)>=2 and Path(tokens[0]).name=="deltawire" and tokens[1] in {"validate","render","check"}:
+   found.append({"command":tokens[1],"invocation":segment.strip(),"shell_exit_code":None})
+ if len(found)==1 and len(commands)==1: found[0]["shell_exit_code"]=observation["shell_exit_code"]
+ elif found and observation["shell_exit_code"]==0 and operators and all(op=="&&" for op in operators):
+  for item in found:item["shell_exit_code"]=0
+ return found
+
+def invokes_deltawire(observation):
+ for segment in re.split(r"&&|\|\||;|\n",observation["command"]):
+  try: tokens=shlex.split(segment)
+  except ValueError: continue
+  while tokens and ("=" in tokens[0] and not tokens[0].startswith("/")): tokens=tokens[1:]
+  if tokens and tokens[0] in {"sudo","env"}: tokens=tokens[1:]
+  if tokens and Path(tokens[0]).name=="deltawire": return True
+ return False
+
+def parse(path):
+ observations=[observation(c) for c in calls(path) if c["name"] in SHELL_NAMES]
+ return observations,[item for obs in observations for item in lifecycle(obs)]
diff --git a/labs/20-deltawire/eval/scripts/v4/test_environment_receipt.py b/labs/20-deltawire/eval/scripts/v4/test_environment_receipt.py
new file mode 100644
index 000000000..7604c8193
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/test_environment_receipt.py
@@ -0,0 +1,26 @@
+#!/usr/bin/env python3
+import hashlib,json,os,subprocess,tempfile
+from pathlib import Path
+HERE=Path(__file__).resolve().parent
+def sha(p):return hashlib.sha256(p.read_bytes()).hexdigest()
+def run(root,binary,mutate=None):
+ files={}
+ for name in ("config","schema","public_contract"):
+  p=root/f"{name}.json";p.write_text(name);files[name]={"path":str(p),"sha256":sha(p)}
+ exp={"binary":{"path":str(binary),"realpath":str(binary.resolve()),"sha256":sha(binary) if binary.is_file() else "0"*64,"version":"deltawire version dev"},"files":files}
+ if mutate:mutate(exp)
+ ep=root/"expect.json";rp=root/"receipt.json";ep.write_text(json.dumps(exp));result=subprocess.run(["python3",str(HERE/"environment_receipt.py"),"--expectations",str(ep),"--receipt",str(rp)],capture_output=True,text=True);return result.returncode,json.loads(rp.read_text())
+def main():
+ with tempfile.TemporaryDirectory() as tmp:
+  root=Path(tmp);binary=root/"deltawire";binary.write_text("#!/bin/sh\nprintf 'deltawire version dev\\nsupported plan version deltawire.plan.v1\\n'\n");binary.chmod(0o755)
+  assert run(root,binary)[1]["status"]=="pass"
+  assert run(root,binary,lambda e:e["binary"].update(sha256="0"*64))[1]["status"]=="fail"
+  assert run(root,binary,lambda e:e["binary"].update(path=str(root/"wrong")))[1]["status"]=="fail"
+  assert run(root,binary,lambda e:e["binary"].update(version="wrong"))[1]["status"]=="fail"
+  assert run(root,binary,lambda e:e["files"]["config"].update(sha256="0"*64))[1]["status"]=="fail"
+  assert run(root,binary,lambda e:e["binary"].update(realpath="/wrong"))[1]["status"]=="fail"
+  binary.chmod(0o644);assert run(root,binary)[1]["status"]=="fail";binary.chmod(0o755)
+  link=root/"link";link.symlink_to(binary);assert run(root,link)[1]["status"]=="fail"
+  missing=root/"missing";code,receipt=run(root,missing);assert code and receipt["status"]=="fail"
+ print("Deterministic environment receipt tests passed.")
+if __name__=="__main__":main()
diff --git a/labs/20-deltawire/eval/scripts/v4/test_fidelity.py b/labs/20-deltawire/eval/scripts/v4/test_fidelity.py
new file mode 100644
index 000000000..7fb11453e
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/test_fidelity.py
@@ -0,0 +1,29 @@
+#!/usr/bin/env python3
+import json,tempfile
+from pathlib import Path
+from fidelity import build
+EVAL=Path(__file__).resolve().parents[2];SPEC=json.loads((EVAL/"tasks/range-large/task-spec.json").read_text())
+def event(name,args,output="",status="success"):
+ return {"toolCalls":[{"name":name,"args":args,"status":status,"result":[{"functionResponse":{"response":{"output":output}}}]}]}
+GOOD=[event("run_shell_command",{"command":"gemini skills list"},"deltawire"),event("activate_skill",{"name":"deltawire"}),event("read_file",{"file_path":"/task-contract.json"}),event("run_shell_command",{"command":"deltawire validate --repo . p"}),event("run_shell_command",{"command":"deltawire render --repo . p"}),event("run_shell_command",{"command":"deltawire check --repo . p"})]
+def run(events,arm="D1",state=True,plan=True):
+ with tempfile.TemporaryDirectory() as tmp:
+  root=Path(tmp);trajectory=root/"trajectory.jsonl";trajectory.write_text("\n".join(json.dumps(e) for e in events));art=root/"artifacts"
+  paths=[art/".deltawire/config.json",art/".deltawire/schemas/range-large.schema.json",art/"testdata/generated/output.ndjson"]
+  if plan:paths += [art/".deltawire/plans/range-large.dw.json",art/".deltawire/plan-contract-receipt.json"]
+  if state:paths.append(art/".deltawire/state.json")
+  for p in paths:p.parent.mkdir(parents=True,exist_ok=True);p.write_text("{}\n")
+  (art/".deltawire/schemas/range-large.schema.json").write_bytes((EVAL/"tasks/range-large/authoritative-schema.json").read_bytes())
+  if plan:(art/".deltawire/plan-contract-receipt.json").write_text('{"status":"pass"}\n')
+  ind=root/"ind.json";ind.write_text('{"status":"pass"}\n')
+  return build(trajectory,root,arm,SPEC,1,ind)
+def main():
+ assert run(GOOD)["status"]=="pass"
+ for i in range(len(GOOD)):assert run(GOOD[:i]+GOOD[i+1:])["status"]=="fail"
+ assert run(GOOD,state=False)["status"]=="fail"
+ failed=GOOD[:-3]+[event("run_shell_command",{"command":"deltawire validate p"},"Exit Code: 2\nProcess Group PGID: 1"),*GOOD[-2:]];assert run(failed)["status"]=="fail"
+ assert run([],arm="B0",state=False,plan=False)["status"]=="pass"
+ assert run([],arm="B0",state=False,plan=True)["status"]=="fail"
+ assert run([event("run_shell_command",{"command":"/usr/local/bin/deltawire version"})],arm="B0",state=False,plan=False)["status"]=="fail"
+ print("Treatment-use v4 tests passed.")
+if __name__=="__main__":main()
diff --git a/labs/20-deltawire/eval/scripts/v4/test_runner.py b/labs/20-deltawire/eval/scripts/v4/test_runner.py
new file mode 100644
index 000000000..dba9ca86d
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/test_runner.py
@@ -0,0 +1,16 @@
+#!/usr/bin/env python3
+import tempfile
+from pathlib import Path
+import runner
+def expect(text,fn,*args):
+ try:fn(*args);raise AssertionError("accepted invalid state")
+ except RuntimeError as e:assert text in str(e),str(e)
+def main():
+ runs=runner.frozen_pair();assert [x["arm"] for x in runs]==["D1","B0"]
+ with tempfile.TemporaryDirectory() as tmp:
+  root=Path(tmp);runner.refuse(root,"v4");(root/"partial").write_text("x");expect("rerun forbidden",runner.refuse,root,"v4")
+ expect("probe-release.v2 missing",runner.verify_release,"missing")
+ source=(Path(__file__).parent/"runner.py").read_text();assert 'choices=("probe-v4","pair")' in source
+ for forbidden in ('"full"','probe-v3','probe-v2','allow-72'):assert forbidden not in source
+ print("V4 runner authorization, adjacency, release, and rerun gates passed.")
+if __name__=="__main__":main()
diff --git a/labs/20-deltawire/eval/scripts/v4/test_shell_observation.py b/labs/20-deltawire/eval/scripts/v4/test_shell_observation.py
new file mode 100644
index 000000000..de3ba9bfe
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/test_shell_observation.py
@@ -0,0 +1,19 @@
+#!/usr/bin/env python3
+import json,tempfile
+from pathlib import Path
+from shell_observation import parse
+def event(command,output="",status="success",extra=None):
+ response={"output":output};response.update(extra or {})
+ return {"toolCalls":[{"name":"run_shell_command","args":{"command":command},"status":status,"result":[{"functionResponse":{"response":response}}]}]}
+def run(events):
+ with tempfile.TemporaryDirectory() as tmp:
+  p=Path(tmp)/"t.jsonl";p.write_text("\n".join(json.dumps(e) for e in events));return parse(p)
+def main():
+ obs,life=run([event("deltawire validate p","ok\nExit Code: 2\nProcess Group PGID: 7\n</untrusted_context>")]);assert obs[0]["tool_status"]=="success" and obs[0]["shell_exit_code"]==2 and life[0]["shell_exit_code"]==2
+ obs,_=run([event("which deltawire","/usr/local/bin/deltawire")]);assert obs[0]["shell_exit_code"] is None
+ obs,_=run([event("type -P deltawire","/fake\nExit Code: 0\nProcess Group PGID: 1",extra={"exit_code":3})]);assert obs[0]["shell_exit_code"]==3 and obs[0]["shell_exit_provenance"]=="structured"
+ _,life=run([event("/usr/local/bin/deltawire validate p && deltawire render p && deltawire check p","Exit Code: 0\nProcess Group PGID: 9")]);assert [x["command"] for x in life]==["validate","render","check"] and all(x["shell_exit_code"]==0 for x in life)
+ _,life=run([event("deltawire validate p; deltawire check p","Exit Code: 0\nProcess Group PGID: 9")]);assert all(x["shell_exit_code"] is None for x in life)
+ _,life=run([event("echo 'deltawire validate'",status="success")]);assert not life
+ print("Canonical shell observation tests passed.")
+if __name__=="__main__":main()
diff --git a/labs/20-deltawire/eval/scripts/v4/validate.py b/labs/20-deltawire/eval/scripts/v4/validate.py
new file mode 100644
index 000000000..7e68e8379
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/validate.py
@@ -0,0 +1,23 @@
+#!/usr/bin/env python3
+import hashlib,json
+from pathlib import Path
+ROOT=Path(__file__).resolve().parents[5];EVAL=ROOT/"labs/20-deltawire/eval";V4=EVAL/"probes/range-large-v4"
+def sha(p):return hashlib.sha256(p.read_bytes()).hexdigest()
+def main():
+ canonical=EVAL/"tasks/range-large";pairs=[V4,V4/"pair-task"]
+ for base in pairs:
+  for name in ("task-spec.json","authoritative-schema.json"):
+   assert (base/name).read_bytes()==(canonical/name).read_bytes(),f"v4 projection mismatch: {base/name}"
+  assert (base/"environment/task-contract.json").read_bytes()==(canonical/"task-spec.json").read_bytes()
+  assert (base/"tests/task-spec.json").read_bytes()==(canonical/"task-spec.json").read_bytes()
+  artifacts=(base/"task.toml").read_text()
+  for required in ("state.json","environment-receipt.json","plan-contract-receipt.json","output.ndjson"):assert required in artifacts
+  assert (base/"environment/environment_receipt.py").read_bytes()==(EVAL/"scripts/v4/environment_receipt.py").read_bytes()
+ matrix=json.loads((EVAL/"matrices/72-run-matrix.json").read_text());assert [matrix[i]["arm"] for i in (36,37)]==["D1","B0"] and all(matrix[i]["pair_id"]=="range-large/r1" for i in (36,37))
+ workflow=(ROOT/".github/workflows/deltawire-preflight-v4.yml").read_text();assert "GEMINI_API_KEY" not in workflow and "harbor run" not in workflow
+ manifest=EVAL/"manifests/preflight-v4-range-large.json"
+ if manifest.is_file():
+  data=json.loads(manifest.read_text());assert data["ready_for_72"] is False and data["matrix_positions"]["zero_based"]==[36,37]
+  for name,want in data["input_hashes"].items():assert (ROOT/name).is_file() and sha(ROOT/name)==want,name
+ print("V4 projections, artifacts, matrix positions, manifest, and no-live-CI checks passed.")
+if __name__=="__main__":main()
diff --git a/labs/20-deltawire/eval/scripts/v4/validate.sh b/labs/20-deltawire/eval/scripts/v4/validate.sh
new file mode 100644
index 000000000..286e5e9b5
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/validate.sh
@@ -0,0 +1,13 @@
+#!/usr/bin/env bash
+set -euo pipefail
+ROOT="$(cd "$(dirname "$0")/../../../../.." && pwd)";cd "$ROOT"
+python3 labs/20-deltawire/eval/scripts/v4/verify_evidence_locks.py
+python3 labs/20-deltawire/eval/scripts/v4/test_environment_receipt.py
+python3 labs/20-deltawire/eval/scripts/v4/test_shell_observation.py
+python3 labs/20-deltawire/eval/scripts/v4/test_fidelity.py
+python3 labs/20-deltawire/eval/scripts/v4/test_runner.py
+SHA="$(bash labs/20-deltawire/eval/scripts/v4/build_and_stage.sh)"
+test "$SHA" = "$(python3 -c 'import json;print(json.load(open("labs/20-deltawire/eval/probes/range-large-v4/environment/deltawire-environment-expectations.json"))["binary"]["sha256"])')"
+python3 labs/20-deltawire/eval/scripts/v4/validate.py
+python3 -m py_compile labs/20-deltawire/eval/scripts/v4/*.py
+git diff --check
diff --git a/labs/20-deltawire/eval/scripts/v4/verify_evidence_locks.py b/labs/20-deltawire/eval/scripts/v4/verify_evidence_locks.py
new file mode 100644
index 000000000..9a5eb61b1
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/verify_evidence_locks.py
@@ -0,0 +1,20 @@
+#!/usr/bin/env python3
+import hashlib,json,subprocess
+from pathlib import Path
+ROOT=Path(__file__).resolve().parents[5]; EVAL=ROOT/"labs/20-deltawire/eval"
+
+def sha(data): return hashlib.sha256(data).hexdigest()
+def main():
+ for path in (EVAL/"results/preflight-v1/treatment-probe-range-large/v1-evidence-lock.json",EVAL/"manifests/preflight-v2-evidence-lock.json"):
+  lock=json.loads(path.read_text())
+  for name,want in lock["immutable_files"].items():
+   target=ROOT/name
+   if not target.is_file() or sha(target.read_bytes())!=want: raise SystemExit(f"evidence mismatch: {name}")
+ lock=json.loads((EVAL/"manifests/preflight-v3-evidence-lock.json").read_text()); commit=lock["source_commit"]
+ tree=subprocess.check_output(["git","rev-parse",f"{commit}^{{tree}}"],cwd=ROOT,text=True).strip()
+ if tree!=lock["source_tree"]: raise SystemExit("v3 source tree mismatch")
+ for name,want in lock["immutable_files"].items():
+  data=subprocess.check_output(["git","show",f"{commit}:{name}"],cwd=ROOT)
+  if sha(data)!=want: raise SystemExit(f"v3 Git-object mismatch: {name}")
+ print(f"v1-v3 evidence locks passed ({len(lock['immutable_files'])} v3 files).")
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/v4/verify_plan_contract.py b/labs/20-deltawire/eval/scripts/v4/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/v4/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/validate_dataset.sh b/labs/20-deltawire/eval/scripts/validate_dataset.sh
index 9d16271b0..8db1060b4 100755
--- a/labs/20-deltawire/eval/scripts/validate_dataset.sh
+++ b/labs/20-deltawire/eval/scripts/validate_dataset.sh
@@ -1,2 +1,20 @@
-echo 'validate_dataset.sh stub'
-exit 0
\ No newline at end of file
+#!/usr/bin/env bash
+set -euo pipefail
+ROOT="$(cd "$(dirname "$0")/../../../.." && pwd)"
+cd "$ROOT"
+python3 labs/20-deltawire/eval/scripts/validate_skill.py labs/20-deltawire/eval/skills/deltawire/SKILL.md
+python3 labs/20-deltawire/eval/scripts/task_contracts.py check
+python3 labs/20-deltawire/eval/scripts/test_task_contracts.py
+python3 labs/20-deltawire/eval/scripts/test_plan_contract.py
+python3 labs/20-deltawire/eval/scripts/test_fidelity.py
+python3 labs/20-deltawire/eval/scripts/test_runner_gates.py
+STAGED_SHA="$(bash labs/20-deltawire/eval/scripts/build_and_manifest.sh)"
+python3 - "$STAGED_SHA" <<'PY'
+import hashlib,pathlib,sys
+root=pathlib.Path("labs/20-deltawire/eval"); expected=sys.argv[1]
+paths=[root/".generated/deltawire-linux-amd64"]+sorted(root.glob("tasks/*/environment/.generated/deltawire"))+sorted(root.glob("probes/*/environment/.generated/deltawire"))
+assert paths and all(hashlib.sha256(path.read_bytes()).hexdigest()==expected for path in paths)
+print(f"Identical Linux binary staged into {len(paths)-1} task contexts.")
+PY
+python3 labs/20-deltawire/eval/scripts/validate_preflight.py
+bash labs/20-deltawire/scripts/check-production-cleanliness.sh
diff --git a/labs/20-deltawire/eval/scripts/validate_preflight.py b/labs/20-deltawire/eval/scripts/validate_preflight.py
new file mode 100644
index 000000000..400280dd3
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/validate_preflight.py
@@ -0,0 +1,45 @@
+#!/usr/bin/env python3
+import hashlib
+import json
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[4]
+EVAL = ROOT / "labs/20-deltawire/eval"
+
+
+def main():
+    matrix = json.loads((EVAL / "matrices/72-run-matrix.json").read_text(encoding="utf-8"))
+    assert len(matrix) == 72
+    assert len({(row["pair_id"], row["arm"]) for row in matrix}) == 72
+    assert all(matrix[i]["pair_id"] == matrix[i+1]["pair_id"] and {matrix[i]["arm"], matrix[i+1]["arm"]} == {"B0", "D1"} for i in range(0, 72, 2))
+    assert all(not Path(value).is_absolute() for row in matrix for value in row.values() if isinstance(value, str))
+    projection_path = EVAL / "results/preflight-v4/integration-projection.json"
+    projection = json.loads(projection_path.read_text(encoding="utf-8"))
+    assert projection["schema_version"] == "deltawire-integration-projection.v1"
+    assert projection["evidence_commit"] == "c200dc0eb3c4dff0e4732cc4b2daa1cd535703e3"
+    assert projection["excluded_path_segments"] == ["/raw/"]
+    for lock_path in (EVAL/"results/preflight-v1/treatment-probe-range-large/v1-evidence-lock.json", EVAL/"manifests/preflight-v2-evidence-lock.json"):
+        lock=json.loads(lock_path.read_text())
+        for relative,expected in lock["immutable_files"].items():
+            path=ROOT/relative
+            if not path.is_file():
+                if "/raw/" in relative:
+                    continue
+                raise AssertionError(f"locked compact evidence missing: {relative}")
+            if hashlib.sha256(path.read_bytes()).hexdigest()!=expected: raise AssertionError(f"locked evidence changed: {relative}")
+    runner=(EVAL/"scripts/runner.py").read_text()
+    assert 'choices=("probe-v3", "pair", "full")' in runner
+    assert 'probe-v2' not in runner and 'probe-v1' not in runner
+    assert 'full mode is unconditionally unauthorized' in runner
+    v2=EVAL/"probes/range-large-v2"; v3=EVAL/"probes/range-large-v3"
+    assert v2.is_dir() and v3.is_dir() and v2.resolve()!=v3.resolve()
+    assert 'operatorstack/range-large-treatment-probe-v3' in (v3/"task.toml").read_text()
+    assert (v2/"task.toml").read_bytes() != (v3/"task.toml").read_bytes()
+    assert 'results/preflight-v3/treatment-probe-range-large-v3' in runner
+    workflow=(ROOT/".github/workflows/deltawire-lab.yml").read_text(); assert "GEMINI_API_KEY" not in workflow and "harbor run" not in workflow
+    readiness_files=list((EVAL/"results").glob("preflight-v*/readiness.json"))
+    assert readiness_files and all(json.loads(path.read_text()).get("READY_FOR_72") is False for path in readiness_files)
+    print("Frozen matrix, v1/v2 immutability, v3 isolation, and READY_FOR_72=false passed.")
+
+
+if __name__ == "__main__": main()
diff --git a/labs/20-deltawire/eval/scripts/validate_skill.py b/labs/20-deltawire/eval/scripts/validate_skill.py
new file mode 100755
index 000000000..46c488869
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/validate_skill.py
@@ -0,0 +1,59 @@
+#!/usr/bin/env python3
+import re
+import sys
+from pathlib import Path
+
+
+REQUIRED_BODY = {
+    "discovery": "gemini skills list",
+    "availability": "command -v deltawire",
+    "version": "deltawire version",
+    "applicability": "compact deterministic description",
+    "abstention": "too irregular",
+    "public contract": "/task-contract.json",
+    "immutable schema": "schema is immutable",
+    "plan": ".deltawire/plans/",
+    "validate": "deltawire validate",
+    "contract check": "verify_plan_contract.py",
+    "render": "deltawire render",
+    "check": "deltawire check",
+    "manual generation prohibition": "manual expanded-data generation",
+    "schema-edit prohibition": "schema edits",
+    "oracle": "semantic oracle",
+}
+
+
+def validate(path: Path) -> list[str]:
+    text = path.read_text(encoding="utf-8") if path.is_file() else ""
+    errors = []
+    match = re.match(r"\A---\n(.*?)\n---\n(.*)\Z", text, re.DOTALL)
+    if not match:
+        return ["missing or invalid frontmatter"]
+    frontmatter, body = match.groups()
+    fields = {}
+    for line in frontmatter.splitlines():
+        if ":" in line:
+            key, value = line.split(":", 1)
+            fields[key.strip()] = value.strip()
+    if not fields.get("name"):
+        errors.append("missing name")
+    if not fields.get("description"):
+        errors.append("missing description")
+    elif "gemini skills list" not in fields["description"] or "before activating" not in fields["description"].lower():
+        errors.append("frontmatter must require pre-activation skill discovery")
+    if not body.strip():
+        errors.append("empty body")
+    lower = body.lower()
+    for label, guidance in REQUIRED_BODY.items():
+        if guidance not in lower:
+            errors.append(f"missing {label} guidance")
+    return errors
+
+
+if __name__ == "__main__":
+    target = Path(sys.argv[1])
+    problems = validate(target)
+    if problems:
+        print("\n".join(problems), file=sys.stderr)
+        raise SystemExit(1)
+    print("DeltaWire skill validation passed.")
diff --git a/labs/20-deltawire/eval/scripts/verify_plan_contract.py b/labs/20-deltawire/eval/scripts/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/scripts/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/skills/deltawire/SKILL.md b/labs/20-deltawire/eval/skills/deltawire/SKILL.md
index 38ce0f46b..597e9532e 100644
--- a/labs/20-deltawire/eval/skills/deltawire/SKILL.md
+++ b/labs/20-deltawire/eval/skills/deltawire/SKILL.md
@@ -1 +1,33 @@
-Use deltawire to materialize repetitive data.
\ No newline at end of file
+---
+name: deltawire
+description: Before activating, use the shell to run `gemini skills list` and confirm `deltawire` is listed. Then use this skill for deterministic generation of large repetitive JSON or NDJSON datasets from Cartesian products, ranges, repeated rows, variants, or regular-plus-exception patterns. Do not use for very small outputs or highly irregular records.
+---
+
+# DeltaWire
+
+Before activating this skill, discovery must be proven by a successful
+structured shell call to `gemini skills list` whose output lists `deltawire`.
+Before choosing DeltaWire, run `command -v deltawire` and `deltawire version`.
+Use it when a dataset has a compact deterministic description: Cartesian
+products, numeric ranges, repeated rows, systematic variants, or a regular
+base plus explicit exceptions. Abstain for very small outputs and records that
+are too irregular to express compactly; state that reason explicitly.
+
+Read the authoritative public contract at `/task-contract.json`. Inspect the
+preinstalled schema named by `authoritative_schema.path`, and prove its SHA-256
+matches `authoritative_schema.sha256`. The schema is immutable: do not create,
+replace, or edit it. Create only the compact plan under `.deltawire/plans/`,
+binding its schema, output, record count, and assertions to the public contract.
+Also prove the SHA-256 of `/task-contract.json` and `.deltawire/config.json`;
+both are immutable and must remain byte-identical to their preinstalled copies.
+
+Run the complete lifecycle:
+
+1. `deltawire validate --repo . <plan>`
+2. `verify_plan_contract.py --contract /task-contract.json --repo . --plan <plan> --receipt .deltawire/plan-contract-receipt.json`
+3. `deltawire render --repo . <plan>`
+4. `deltawire check --repo . <plan>`
+
+After choosing DeltaWire, manual expanded-data generation, manual output edits,
+schema edits, and access to hidden verifier files are prohibited. The task's
+independent semantic oracle remains the final correctness authority.
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/authoritative-schema.json b/labs/20-deltawire/eval/tasks/irregular-large/authoritative-schema.json
new file mode 100644
index 000000000..ce7e02d94
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-large/authoritative-schema.json
@@ -0,0 +1,23 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "maxProperties": 2,
+  "minProperties": 2,
+  "patternProperties": {
+    "^field_[0-9]+$": {
+      "type": "integer"
+    }
+  },
+  "properties": {
+    "id": {
+      "maximum": 150,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "id"
+  ],
+  "title": "irregular-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/irregular-large/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-large/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/environment/.deltawire/schemas/irregular-large.schema.json b/labs/20-deltawire/eval/tasks/irregular-large/environment/.deltawire/schemas/irregular-large.schema.json
new file mode 100644
index 000000000..ce7e02d94
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-large/environment/.deltawire/schemas/irregular-large.schema.json
@@ -0,0 +1,23 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "maxProperties": 2,
+  "minProperties": 2,
+  "patternProperties": {
+    "^field_[0-9]+$": {
+      "type": "integer"
+    }
+  },
+  "properties": {
+    "id": {
+      "maximum": 150,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "id"
+  ],
+  "title": "irregular-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/environment/Dockerfile b/labs/20-deltawire/eval/tasks/irregular-large/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/irregular-large/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/irregular-large/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/environment/task-contract.json b/labs/20-deltawire/eval/tasks/irregular-large/environment/task-contract.json
new file mode 100644
index 000000000..826ccde01
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-large/environment/task-contract.json
@@ -0,0 +1,25 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/irregular-large.schema.json",
+    "sha256": "fc0f2610d6fd2d7656030388500f91ef465560137ab7c31aa8698bca362bb746"
+  },
+  "deltawire_applicability": "should_abstain",
+  "family": "irregular",
+  "generation": {
+    "dynamic_field_prefix": "field_",
+    "field_order": [
+      "id",
+      "dynamic"
+    ],
+    "id_start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 150,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "irregular-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/irregular-large/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-large/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/instruction.md b/labs/20-deltawire/eval/tasks/irregular-large/instruction.md
index 36586cf78..b81bf4669 100644
--- a/labs/20-deltawire/eval/tasks/irregular-large/instruction.md
+++ b/labs/20-deltawire/eval/tasks/irregular-large/instruction.md
@@ -1 +1,29 @@
-# Instruction
+Generate the exact `irregular-large` dataset defined by public `/task-contract.json`. Write exactly 150 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/irregular-large.schema.json",
+    "sha256": "fc0f2610d6fd2d7656030388500f91ef465560137ab7c31aa8698bca362bb746"
+  },
+  "deltawire_applicability": "should_abstain",
+  "family": "irregular",
+  "generation": {
+    "dynamic_field_prefix": "field_",
+    "field_order": [
+      "id",
+      "dynamic"
+    ],
+    "id_start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 150,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "irregular-large"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/task-spec.json b/labs/20-deltawire/eval/tasks/irregular-large/task-spec.json
new file mode 100644
index 000000000..826ccde01
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-large/task-spec.json
@@ -0,0 +1,25 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/irregular-large.schema.json",
+    "sha256": "fc0f2610d6fd2d7656030388500f91ef465560137ab7c31aa8698bca362bb746"
+  },
+  "deltawire_applicability": "should_abstain",
+  "family": "irregular",
+  "generation": {
+    "dynamic_field_prefix": "field_",
+    "field_order": [
+      "id",
+      "dynamic"
+    ],
+    "id_start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 150,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "irregular-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/task.toml b/labs/20-deltawire/eval/tasks/irregular-large/task.toml
index 156d6dc8a..ac2ee0690 100644
--- a/labs/20-deltawire/eval/tasks/irregular-large/task.toml
+++ b/labs/20-deltawire/eval/tasks/irregular-large/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/irregular-large.dw.json", ".deltawire/schemas/irregular-large.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/irregular-large"
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/irregular-large/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-large/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/tests/task-spec.json b/labs/20-deltawire/eval/tasks/irregular-large/tests/task-spec.json
new file mode 100644
index 000000000..826ccde01
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-large/tests/task-spec.json
@@ -0,0 +1,25 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/irregular-large.schema.json",
+    "sha256": "fc0f2610d6fd2d7656030388500f91ef465560137ab7c31aa8698bca362bb746"
+  },
+  "deltawire_applicability": "should_abstain",
+  "family": "irregular",
+  "generation": {
+    "dynamic_field_prefix": "field_",
+    "field_order": [
+      "id",
+      "dynamic"
+    ],
+    "id_start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 150,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "irregular-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-large/tests/test.sh b/labs/20-deltawire/eval/tasks/irregular-large/tests/test.sh
old mode 100644
new mode 100755
index 06bd98656..9b56450e6
--- a/labs/20-deltawire/eval/tasks/irregular-large/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/irregular-large/tests/test.sh
@@ -1,2 +1,7 @@
 #!/bin/bash
-exit 0
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/authoritative-schema.json b/labs/20-deltawire/eval/tasks/irregular-medium/authoritative-schema.json
new file mode 100644
index 000000000..c6910c1e6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/authoritative-schema.json
@@ -0,0 +1,23 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "maxProperties": 2,
+  "minProperties": 2,
+  "patternProperties": {
+    "^field_[0-9]+$": {
+      "type": "integer"
+    }
+  },
+  "properties": {
+    "id": {
+      "maximum": 15,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "id"
+  ],
+  "title": "irregular-medium",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/irregular-medium/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/environment/.deltawire/schemas/irregular-medium.schema.json b/labs/20-deltawire/eval/tasks/irregular-medium/environment/.deltawire/schemas/irregular-medium.schema.json
new file mode 100644
index 000000000..c6910c1e6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/environment/.deltawire/schemas/irregular-medium.schema.json
@@ -0,0 +1,23 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "maxProperties": 2,
+  "minProperties": 2,
+  "patternProperties": {
+    "^field_[0-9]+$": {
+      "type": "integer"
+    }
+  },
+  "properties": {
+    "id": {
+      "maximum": 15,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "id"
+  ],
+  "title": "irregular-medium",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/environment/Dockerfile b/labs/20-deltawire/eval/tasks/irregular-medium/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/irregular-medium/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/environment/task-contract.json b/labs/20-deltawire/eval/tasks/irregular-medium/environment/task-contract.json
new file mode 100644
index 000000000..0e041d447
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/environment/task-contract.json
@@ -0,0 +1,25 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/irregular-medium.schema.json",
+    "sha256": "da1ffa1df0e6c5680c53122a6463d5d61ca9e90e4b83628dd35ad30c0ba81da6"
+  },
+  "deltawire_applicability": "should_abstain",
+  "family": "irregular",
+  "generation": {
+    "dynamic_field_prefix": "field_",
+    "field_order": [
+      "id",
+      "dynamic"
+    ],
+    "id_start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 15,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "irregular-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/irregular-medium/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/instruction.md b/labs/20-deltawire/eval/tasks/irregular-medium/instruction.md
index 00d1948fd..a79c0d4e7 100644
--- a/labs/20-deltawire/eval/tasks/irregular-medium/instruction.md
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/instruction.md
@@ -1,3 +1,29 @@
-Generate a dataset of 5 unique irregular logging events.
-Output the data to `testdata/generated/irregular-logs.ndjson`.
-Each event should have entirely distinct structures and fields.
+Generate the exact `irregular-medium` dataset defined by public `/task-contract.json`. Write exactly 15 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/irregular-medium.schema.json",
+    "sha256": "da1ffa1df0e6c5680c53122a6463d5d61ca9e90e4b83628dd35ad30c0ba81da6"
+  },
+  "deltawire_applicability": "should_abstain",
+  "family": "irregular",
+  "generation": {
+    "dynamic_field_prefix": "field_",
+    "field_order": [
+      "id",
+      "dynamic"
+    ],
+    "id_start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 15,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "irregular-medium"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/task-spec.json b/labs/20-deltawire/eval/tasks/irregular-medium/task-spec.json
new file mode 100644
index 000000000..0e041d447
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/task-spec.json
@@ -0,0 +1,25 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/irregular-medium.schema.json",
+    "sha256": "da1ffa1df0e6c5680c53122a6463d5d61ca9e90e4b83628dd35ad30c0ba81da6"
+  },
+  "deltawire_applicability": "should_abstain",
+  "family": "irregular",
+  "generation": {
+    "dynamic_field_prefix": "field_",
+    "field_order": [
+      "id",
+      "dynamic"
+    ],
+    "id_start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 15,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "irregular-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/task.toml b/labs/20-deltawire/eval/tasks/irregular-medium/task.toml
index a28b65995..8ad3878f9 100644
--- a/labs/20-deltawire/eval/tasks/irregular-medium/task.toml
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/irregular-medium.dw.json", ".deltawire/schemas/irregular-medium.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/irregular-medium"
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/irregular-medium/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/tests/task-spec.json b/labs/20-deltawire/eval/tasks/irregular-medium/tests/task-spec.json
new file mode 100644
index 000000000..0e041d447
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/tests/task-spec.json
@@ -0,0 +1,25 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/irregular-medium.schema.json",
+    "sha256": "da1ffa1df0e6c5680c53122a6463d5d61ca9e90e4b83628dd35ad30c0ba81da6"
+  },
+  "deltawire_applicability": "should_abstain",
+  "family": "irregular",
+  "generation": {
+    "dynamic_field_prefix": "field_",
+    "field_order": [
+      "id",
+      "dynamic"
+    ],
+    "id_start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 15,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "irregular-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/irregular-medium/tests/test.sh b/labs/20-deltawire/eval/tasks/irregular-medium/tests/test.sh
old mode 100644
new mode 100755
index 46e887a60..9b56450e6
--- a/labs/20-deltawire/eval/tasks/irregular-medium/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/irregular-medium/tests/test.sh
@@ -1,6 +1,7 @@
 #!/bin/bash
-if [ -f testdata/generated/irregular-logs.ndjson ]; then
-    echo '{"exact_match": 1}'
-else
-    echo '{"exact_match": 0}'
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
 fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/authoritative-schema.json b/labs/20-deltawire/eval/tasks/matrix-large/authoritative-schema.json
new file mode 100644
index 000000000..0a63c6df0
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-large/authoritative-schema.json
@@ -0,0 +1,40 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "endpoint": {
+      "enum": [
+        "/api",
+        "/web",
+        "/login"
+      ]
+    },
+    "method": {
+      "enum": [
+        "GET",
+        "POST",
+        "PUT",
+        "DELETE"
+      ]
+    },
+    "role": {
+      "enum": [
+        "guest",
+        "member",
+        "admin",
+        "superuser"
+      ]
+    },
+    "status": {
+      "const": "deny"
+    }
+  },
+  "required": [
+    "role",
+    "method",
+    "endpoint",
+    "status"
+  ],
+  "title": "matrix-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/matrix-large/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-large/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/environment/.deltawire/schemas/matrix-large.schema.json b/labs/20-deltawire/eval/tasks/matrix-large/environment/.deltawire/schemas/matrix-large.schema.json
new file mode 100644
index 000000000..0a63c6df0
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-large/environment/.deltawire/schemas/matrix-large.schema.json
@@ -0,0 +1,40 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "endpoint": {
+      "enum": [
+        "/api",
+        "/web",
+        "/login"
+      ]
+    },
+    "method": {
+      "enum": [
+        "GET",
+        "POST",
+        "PUT",
+        "DELETE"
+      ]
+    },
+    "role": {
+      "enum": [
+        "guest",
+        "member",
+        "admin",
+        "superuser"
+      ]
+    },
+    "status": {
+      "const": "deny"
+    }
+  },
+  "required": [
+    "role",
+    "method",
+    "endpoint",
+    "status"
+  ],
+  "title": "matrix-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/environment/Dockerfile b/labs/20-deltawire/eval/tasks/matrix-large/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/matrix-large/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/matrix-large/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/environment/task-contract.json b/labs/20-deltawire/eval/tasks/matrix-large/environment/task-contract.json
new file mode 100644
index 000000000..bbda603ae
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-large/environment/task-contract.json
@@ -0,0 +1,47 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/matrix-large.schema.json",
+    "sha256": "45dd2f4c92fe00ae8ae1ded9e44fd949222af68fe2de45ff757374e528ba9e5a"
+  },
+  "deltawire_applicability": "expected",
+  "family": "matrix",
+  "generation": {
+    "constants": {
+      "status": "deny"
+    },
+    "dimensions": {
+      "endpoint": [
+        "/api",
+        "/web",
+        "/login"
+      ],
+      "method": [
+        "GET",
+        "POST",
+        "PUT",
+        "DELETE"
+      ],
+      "role": [
+        "guest",
+        "member",
+        "admin",
+        "superuser"
+      ]
+    },
+    "field_order": [
+      "role",
+      "method",
+      "endpoint",
+      "status"
+    ]
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 48,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "matrix-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/matrix-large/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-large/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/instruction.md b/labs/20-deltawire/eval/tasks/matrix-large/instruction.md
index 36586cf78..b3570c835 100644
--- a/labs/20-deltawire/eval/tasks/matrix-large/instruction.md
+++ b/labs/20-deltawire/eval/tasks/matrix-large/instruction.md
@@ -1 +1,51 @@
-# Instruction
+Generate the exact `matrix-large` dataset defined by public `/task-contract.json`. Write exactly 48 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/matrix-large.schema.json",
+    "sha256": "45dd2f4c92fe00ae8ae1ded9e44fd949222af68fe2de45ff757374e528ba9e5a"
+  },
+  "deltawire_applicability": "expected",
+  "family": "matrix",
+  "generation": {
+    "constants": {
+      "status": "deny"
+    },
+    "dimensions": {
+      "endpoint": [
+        "/api",
+        "/web",
+        "/login"
+      ],
+      "method": [
+        "GET",
+        "POST",
+        "PUT",
+        "DELETE"
+      ],
+      "role": [
+        "guest",
+        "member",
+        "admin",
+        "superuser"
+      ]
+    },
+    "field_order": [
+      "role",
+      "method",
+      "endpoint",
+      "status"
+    ]
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 48,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "matrix-large"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/task-spec.json b/labs/20-deltawire/eval/tasks/matrix-large/task-spec.json
new file mode 100644
index 000000000..bbda603ae
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-large/task-spec.json
@@ -0,0 +1,47 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/matrix-large.schema.json",
+    "sha256": "45dd2f4c92fe00ae8ae1ded9e44fd949222af68fe2de45ff757374e528ba9e5a"
+  },
+  "deltawire_applicability": "expected",
+  "family": "matrix",
+  "generation": {
+    "constants": {
+      "status": "deny"
+    },
+    "dimensions": {
+      "endpoint": [
+        "/api",
+        "/web",
+        "/login"
+      ],
+      "method": [
+        "GET",
+        "POST",
+        "PUT",
+        "DELETE"
+      ],
+      "role": [
+        "guest",
+        "member",
+        "admin",
+        "superuser"
+      ]
+    },
+    "field_order": [
+      "role",
+      "method",
+      "endpoint",
+      "status"
+    ]
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 48,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "matrix-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/task.toml b/labs/20-deltawire/eval/tasks/matrix-large/task.toml
index 15bd3a4b0..a97a2c97a 100644
--- a/labs/20-deltawire/eval/tasks/matrix-large/task.toml
+++ b/labs/20-deltawire/eval/tasks/matrix-large/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/matrix-large.dw.json", ".deltawire/schemas/matrix-large.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/matrix-large"
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/matrix-large/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-large/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/tests/task-spec.json b/labs/20-deltawire/eval/tasks/matrix-large/tests/task-spec.json
new file mode 100644
index 000000000..bbda603ae
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-large/tests/task-spec.json
@@ -0,0 +1,47 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/matrix-large.schema.json",
+    "sha256": "45dd2f4c92fe00ae8ae1ded9e44fd949222af68fe2de45ff757374e528ba9e5a"
+  },
+  "deltawire_applicability": "expected",
+  "family": "matrix",
+  "generation": {
+    "constants": {
+      "status": "deny"
+    },
+    "dimensions": {
+      "endpoint": [
+        "/api",
+        "/web",
+        "/login"
+      ],
+      "method": [
+        "GET",
+        "POST",
+        "PUT",
+        "DELETE"
+      ],
+      "role": [
+        "guest",
+        "member",
+        "admin",
+        "superuser"
+      ]
+    },
+    "field_order": [
+      "role",
+      "method",
+      "endpoint",
+      "status"
+    ]
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 48,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "matrix-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-large/tests/test.sh b/labs/20-deltawire/eval/tasks/matrix-large/tests/test.sh
old mode 100644
new mode 100755
index 06bd98656..9b56450e6
--- a/labs/20-deltawire/eval/tasks/matrix-large/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/matrix-large/tests/test.sh
@@ -1,2 +1,7 @@
 #!/bin/bash
-exit 0
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/authoritative-schema.json b/labs/20-deltawire/eval/tasks/matrix-medium/authoritative-schema.json
new file mode 100644
index 000000000..69e27ffb1
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/authoritative-schema.json
@@ -0,0 +1,29 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "method": {
+      "enum": [
+        "GET",
+        "POST"
+      ]
+    },
+    "role": {
+      "enum": [
+        "guest",
+        "member",
+        "admin"
+      ]
+    },
+    "status": {
+      "const": "deny"
+    }
+  },
+  "required": [
+    "role",
+    "method",
+    "status"
+  ],
+  "title": "matrix-medium",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/matrix-medium/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/environment/.deltawire/schemas/matrix-medium.schema.json b/labs/20-deltawire/eval/tasks/matrix-medium/environment/.deltawire/schemas/matrix-medium.schema.json
new file mode 100644
index 000000000..69e27ffb1
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/environment/.deltawire/schemas/matrix-medium.schema.json
@@ -0,0 +1,29 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "method": {
+      "enum": [
+        "GET",
+        "POST"
+      ]
+    },
+    "role": {
+      "enum": [
+        "guest",
+        "member",
+        "admin"
+      ]
+    },
+    "status": {
+      "const": "deny"
+    }
+  },
+  "required": [
+    "role",
+    "method",
+    "status"
+  ],
+  "title": "matrix-medium",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/environment/Dockerfile b/labs/20-deltawire/eval/tasks/matrix-medium/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/matrix-medium/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/environment/task-contract.json b/labs/20-deltawire/eval/tasks/matrix-medium/environment/task-contract.json
new file mode 100644
index 000000000..0c32170c6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/environment/task-contract.json
@@ -0,0 +1,38 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/matrix-medium.schema.json",
+    "sha256": "d04bbc205d14048c2e518e2e4ab80fdbdd6e04770d60f4860e9cd985ba34b04f"
+  },
+  "deltawire_applicability": "optional",
+  "family": "matrix",
+  "generation": {
+    "constants": {
+      "status": "deny"
+    },
+    "dimensions": {
+      "method": [
+        "GET",
+        "POST"
+      ],
+      "role": [
+        "guest",
+        "member",
+        "admin"
+      ]
+    },
+    "field_order": [
+      "role",
+      "method",
+      "status"
+    ]
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 6,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "matrix-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/matrix-medium/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/instruction.md b/labs/20-deltawire/eval/tasks/matrix-medium/instruction.md
index 63d479da6..afec76e98 100644
--- a/labs/20-deltawire/eval/tasks/matrix-medium/instruction.md
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/instruction.md
@@ -1,8 +1,42 @@
-Generate a dataset of API authorization scenarios.
-Output a JSON array of objects to `testdata/generated/auth-cases.ndjson`.
-Each object must have the following keys: `id`, `suite`, `input`, and `expected`.
-The `suite` is always "auth".
-The `input` is an object with `role` and `route`.
-Generate all combinations of roles: "guest", "member" against routes: "/billing", "/admin".
-For all combinations, `expected` should be "deny".
-Ensure the `id` is formatted as "auth/<role><route>".
+Generate the exact `matrix-medium` dataset defined by public `/task-contract.json`. Write exactly 6 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/matrix-medium.schema.json",
+    "sha256": "d04bbc205d14048c2e518e2e4ab80fdbdd6e04770d60f4860e9cd985ba34b04f"
+  },
+  "deltawire_applicability": "optional",
+  "family": "matrix",
+  "generation": {
+    "constants": {
+      "status": "deny"
+    },
+    "dimensions": {
+      "method": [
+        "GET",
+        "POST"
+      ],
+      "role": [
+        "guest",
+        "member",
+        "admin"
+      ]
+    },
+    "field_order": [
+      "role",
+      "method",
+      "status"
+    ]
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 6,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "matrix-medium"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/task-spec.json b/labs/20-deltawire/eval/tasks/matrix-medium/task-spec.json
new file mode 100644
index 000000000..0c32170c6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/task-spec.json
@@ -0,0 +1,38 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/matrix-medium.schema.json",
+    "sha256": "d04bbc205d14048c2e518e2e4ab80fdbdd6e04770d60f4860e9cd985ba34b04f"
+  },
+  "deltawire_applicability": "optional",
+  "family": "matrix",
+  "generation": {
+    "constants": {
+      "status": "deny"
+    },
+    "dimensions": {
+      "method": [
+        "GET",
+        "POST"
+      ],
+      "role": [
+        "guest",
+        "member",
+        "admin"
+      ]
+    },
+    "field_order": [
+      "role",
+      "method",
+      "status"
+    ]
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 6,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "matrix-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/task.toml b/labs/20-deltawire/eval/tasks/matrix-medium/task.toml
index 66a9b6ec5..1b4c71169 100644
--- a/labs/20-deltawire/eval/tasks/matrix-medium/task.toml
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/matrix-medium.dw.json", ".deltawire/schemas/matrix-medium.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/matrix-medium"
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/matrix-medium/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/tests/task-spec.json b/labs/20-deltawire/eval/tasks/matrix-medium/tests/task-spec.json
new file mode 100644
index 000000000..0c32170c6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/tests/task-spec.json
@@ -0,0 +1,38 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/matrix-medium.schema.json",
+    "sha256": "d04bbc205d14048c2e518e2e4ab80fdbdd6e04770d60f4860e9cd985ba34b04f"
+  },
+  "deltawire_applicability": "optional",
+  "family": "matrix",
+  "generation": {
+    "constants": {
+      "status": "deny"
+    },
+    "dimensions": {
+      "method": [
+        "GET",
+        "POST"
+      ],
+      "role": [
+        "guest",
+        "member",
+        "admin"
+      ]
+    },
+    "field_order": [
+      "role",
+      "method",
+      "status"
+    ]
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 6,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "matrix-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/matrix-medium/tests/test.sh b/labs/20-deltawire/eval/tasks/matrix-medium/tests/test.sh
old mode 100644
new mode 100755
index 26049a463..9b56450e6
--- a/labs/20-deltawire/eval/tasks/matrix-medium/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/matrix-medium/tests/test.sh
@@ -1,6 +1,7 @@
 #!/bin/bash
-if [ -f testdata/generated/auth-cases.ndjson ]; then
-    echo '{"exact_match": 1}'
-else
-    echo '{"exact_match": 0}'
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
 fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/authoritative-schema.json b/labs/20-deltawire/eval/tasks/mixed-large/authoritative-schema.json
new file mode 100644
index 000000000..c64fc6300
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-large/authoritative-schema.json
@@ -0,0 +1,42 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "anyOf": [
+    {
+      "additionalProperties": false,
+      "properties": {
+        "id": {
+          "type": "integer"
+        },
+        "type": {
+          "const": "std"
+        }
+      },
+      "required": [
+        "id",
+        "type"
+      ],
+      "type": "object"
+    },
+    {
+      "additionalProperties": false,
+      "properties": {
+        "error_code": {
+          "const": 500
+        },
+        "id": {
+          "type": "integer"
+        },
+        "type": {
+          "const": "exc"
+        }
+      },
+      "required": [
+        "id",
+        "type",
+        "error_code"
+      ],
+      "type": "object"
+    }
+  ],
+  "title": "mixed-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/mixed-large/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-large/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/environment/.deltawire/schemas/mixed-large.schema.json b/labs/20-deltawire/eval/tasks/mixed-large/environment/.deltawire/schemas/mixed-large.schema.json
new file mode 100644
index 000000000..c64fc6300
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-large/environment/.deltawire/schemas/mixed-large.schema.json
@@ -0,0 +1,42 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "anyOf": [
+    {
+      "additionalProperties": false,
+      "properties": {
+        "id": {
+          "type": "integer"
+        },
+        "type": {
+          "const": "std"
+        }
+      },
+      "required": [
+        "id",
+        "type"
+      ],
+      "type": "object"
+    },
+    {
+      "additionalProperties": false,
+      "properties": {
+        "error_code": {
+          "const": 500
+        },
+        "id": {
+          "type": "integer"
+        },
+        "type": {
+          "const": "exc"
+        }
+      },
+      "required": [
+        "id",
+        "type",
+        "error_code"
+      ],
+      "type": "object"
+    }
+  ],
+  "title": "mixed-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/environment/Dockerfile b/labs/20-deltawire/eval/tasks/mixed-large/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/mixed-large/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/mixed-large/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/environment/task-contract.json b/labs/20-deltawire/eval/tasks/mixed-large/environment/task-contract.json
new file mode 100644
index 000000000..15e649456
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-large/environment/task-contract.json
@@ -0,0 +1,31 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/mixed-large.schema.json",
+    "sha256": "fc0e60434c88d423663fa2aad8937d88c07f3e133dea0ebf8ae52626d81b3dc1"
+  },
+  "deltawire_applicability": "expected",
+  "family": "mixed",
+  "generation": {
+    "error_code": 500,
+    "exception_count": 50,
+    "exception_field_order": [
+      "id",
+      "type",
+      "error_code"
+    ],
+    "field_order": [
+      "id",
+      "type"
+    ],
+    "standard_count": 200
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 250,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "mixed-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/mixed-large/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-large/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/instruction.md b/labs/20-deltawire/eval/tasks/mixed-large/instruction.md
index 36586cf78..4a569e501 100644
--- a/labs/20-deltawire/eval/tasks/mixed-large/instruction.md
+++ b/labs/20-deltawire/eval/tasks/mixed-large/instruction.md
@@ -1 +1,35 @@
-# Instruction
+Generate the exact `mixed-large` dataset defined by public `/task-contract.json`. Write exactly 250 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/mixed-large.schema.json",
+    "sha256": "fc0e60434c88d423663fa2aad8937d88c07f3e133dea0ebf8ae52626d81b3dc1"
+  },
+  "deltawire_applicability": "expected",
+  "family": "mixed",
+  "generation": {
+    "error_code": 500,
+    "exception_count": 50,
+    "exception_field_order": [
+      "id",
+      "type",
+      "error_code"
+    ],
+    "field_order": [
+      "id",
+      "type"
+    ],
+    "standard_count": 200
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 250,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "mixed-large"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/task-spec.json b/labs/20-deltawire/eval/tasks/mixed-large/task-spec.json
new file mode 100644
index 000000000..15e649456
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-large/task-spec.json
@@ -0,0 +1,31 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/mixed-large.schema.json",
+    "sha256": "fc0e60434c88d423663fa2aad8937d88c07f3e133dea0ebf8ae52626d81b3dc1"
+  },
+  "deltawire_applicability": "expected",
+  "family": "mixed",
+  "generation": {
+    "error_code": 500,
+    "exception_count": 50,
+    "exception_field_order": [
+      "id",
+      "type",
+      "error_code"
+    ],
+    "field_order": [
+      "id",
+      "type"
+    ],
+    "standard_count": 200
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 250,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "mixed-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/task.toml b/labs/20-deltawire/eval/tasks/mixed-large/task.toml
index c627664da..01ca2823d 100644
--- a/labs/20-deltawire/eval/tasks/mixed-large/task.toml
+++ b/labs/20-deltawire/eval/tasks/mixed-large/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/mixed-large.dw.json", ".deltawire/schemas/mixed-large.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/mixed-large"
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/mixed-large/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-large/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/tests/task-spec.json b/labs/20-deltawire/eval/tasks/mixed-large/tests/task-spec.json
new file mode 100644
index 000000000..15e649456
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-large/tests/task-spec.json
@@ -0,0 +1,31 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/mixed-large.schema.json",
+    "sha256": "fc0e60434c88d423663fa2aad8937d88c07f3e133dea0ebf8ae52626d81b3dc1"
+  },
+  "deltawire_applicability": "expected",
+  "family": "mixed",
+  "generation": {
+    "error_code": 500,
+    "exception_count": 50,
+    "exception_field_order": [
+      "id",
+      "type",
+      "error_code"
+    ],
+    "field_order": [
+      "id",
+      "type"
+    ],
+    "standard_count": 200
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 250,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "mixed-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-large/tests/test.sh b/labs/20-deltawire/eval/tasks/mixed-large/tests/test.sh
old mode 100644
new mode 100755
index 06bd98656..9b56450e6
--- a/labs/20-deltawire/eval/tasks/mixed-large/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/mixed-large/tests/test.sh
@@ -1,2 +1,7 @@
 #!/bin/bash
-exit 0
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/authoritative-schema.json b/labs/20-deltawire/eval/tasks/mixed-medium/authoritative-schema.json
new file mode 100644
index 000000000..aacd7250b
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/authoritative-schema.json
@@ -0,0 +1,42 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "anyOf": [
+    {
+      "additionalProperties": false,
+      "properties": {
+        "id": {
+          "type": "integer"
+        },
+        "type": {
+          "const": "std"
+        }
+      },
+      "required": [
+        "id",
+        "type"
+      ],
+      "type": "object"
+    },
+    {
+      "additionalProperties": false,
+      "properties": {
+        "error_code": {
+          "const": 500
+        },
+        "id": {
+          "type": "integer"
+        },
+        "type": {
+          "const": "exc"
+        }
+      },
+      "required": [
+        "id",
+        "type",
+        "error_code"
+      ],
+      "type": "object"
+    }
+  ],
+  "title": "mixed-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/mixed-medium/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/environment/.deltawire/schemas/mixed-medium.schema.json b/labs/20-deltawire/eval/tasks/mixed-medium/environment/.deltawire/schemas/mixed-medium.schema.json
new file mode 100644
index 000000000..aacd7250b
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/environment/.deltawire/schemas/mixed-medium.schema.json
@@ -0,0 +1,42 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "anyOf": [
+    {
+      "additionalProperties": false,
+      "properties": {
+        "id": {
+          "type": "integer"
+        },
+        "type": {
+          "const": "std"
+        }
+      },
+      "required": [
+        "id",
+        "type"
+      ],
+      "type": "object"
+    },
+    {
+      "additionalProperties": false,
+      "properties": {
+        "error_code": {
+          "const": 500
+        },
+        "id": {
+          "type": "integer"
+        },
+        "type": {
+          "const": "exc"
+        }
+      },
+      "required": [
+        "id",
+        "type",
+        "error_code"
+      ],
+      "type": "object"
+    }
+  ],
+  "title": "mixed-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/environment/Dockerfile b/labs/20-deltawire/eval/tasks/mixed-medium/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/mixed-medium/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/environment/task-contract.json b/labs/20-deltawire/eval/tasks/mixed-medium/environment/task-contract.json
new file mode 100644
index 000000000..0306588a8
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/environment/task-contract.json
@@ -0,0 +1,31 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/mixed-medium.schema.json",
+    "sha256": "85a8c052a2295ccc1c480aa997439ae60cb93bdf354e71b7ee4bbe4dbf33f478"
+  },
+  "deltawire_applicability": "optional",
+  "family": "mixed",
+  "generation": {
+    "error_code": 500,
+    "exception_count": 5,
+    "exception_field_order": [
+      "id",
+      "type",
+      "error_code"
+    ],
+    "field_order": [
+      "id",
+      "type"
+    ],
+    "standard_count": 20
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 25,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "mixed-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/mixed-medium/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/instruction.md b/labs/20-deltawire/eval/tasks/mixed-medium/instruction.md
index 1b1fc7599..b049e8f50 100644
--- a/labs/20-deltawire/eval/tasks/mixed-medium/instruction.md
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/instruction.md
@@ -1,3 +1,35 @@
-Generate a mixed dataset of configuration parameters and edge cases.
-Output an NDJSON file to `testdata/generated/config-cases.ndjson`.
-Include typical configurations for "prod" and "dev" environments, as well as edge cases with missing or empty parameters.
+Generate the exact `mixed-medium` dataset defined by public `/task-contract.json`. Write exactly 25 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/mixed-medium.schema.json",
+    "sha256": "85a8c052a2295ccc1c480aa997439ae60cb93bdf354e71b7ee4bbe4dbf33f478"
+  },
+  "deltawire_applicability": "optional",
+  "family": "mixed",
+  "generation": {
+    "error_code": 500,
+    "exception_count": 5,
+    "exception_field_order": [
+      "id",
+      "type",
+      "error_code"
+    ],
+    "field_order": [
+      "id",
+      "type"
+    ],
+    "standard_count": 20
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 25,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "mixed-medium"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/task-spec.json b/labs/20-deltawire/eval/tasks/mixed-medium/task-spec.json
new file mode 100644
index 000000000..0306588a8
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/task-spec.json
@@ -0,0 +1,31 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/mixed-medium.schema.json",
+    "sha256": "85a8c052a2295ccc1c480aa997439ae60cb93bdf354e71b7ee4bbe4dbf33f478"
+  },
+  "deltawire_applicability": "optional",
+  "family": "mixed",
+  "generation": {
+    "error_code": 500,
+    "exception_count": 5,
+    "exception_field_order": [
+      "id",
+      "type",
+      "error_code"
+    ],
+    "field_order": [
+      "id",
+      "type"
+    ],
+    "standard_count": 20
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 25,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "mixed-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/task.toml b/labs/20-deltawire/eval/tasks/mixed-medium/task.toml
index 4d24383c7..243776efc 100644
--- a/labs/20-deltawire/eval/tasks/mixed-medium/task.toml
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/mixed-medium.dw.json", ".deltawire/schemas/mixed-medium.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/mixed-medium"
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/mixed-medium/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/tests/task-spec.json b/labs/20-deltawire/eval/tasks/mixed-medium/tests/task-spec.json
new file mode 100644
index 000000000..0306588a8
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/tests/task-spec.json
@@ -0,0 +1,31 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/mixed-medium.schema.json",
+    "sha256": "85a8c052a2295ccc1c480aa997439ae60cb93bdf354e71b7ee4bbe4dbf33f478"
+  },
+  "deltawire_applicability": "optional",
+  "family": "mixed",
+  "generation": {
+    "error_code": 500,
+    "exception_count": 5,
+    "exception_field_order": [
+      "id",
+      "type",
+      "error_code"
+    ],
+    "field_order": [
+      "id",
+      "type"
+    ],
+    "standard_count": 20
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 25,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "mixed-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/mixed-medium/tests/test.sh b/labs/20-deltawire/eval/tasks/mixed-medium/tests/test.sh
old mode 100644
new mode 100755
index ce5b3f12e..9b56450e6
--- a/labs/20-deltawire/eval/tasks/mixed-medium/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/mixed-medium/tests/test.sh
@@ -1,6 +1,7 @@
 #!/bin/bash
-if [ -f testdata/generated/config-cases.ndjson ]; then
-    echo '{"exact_match": 1}'
-else
-    echo '{"exact_match": 0}'
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
 fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/range-large/authoritative-schema.json b/labs/20-deltawire/eval/tasks/range-large/authoritative-schema.json
new file mode 100644
index 000000000..69b476237
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-large/authoritative-schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 500,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/schemas/range-large.schema.json b/labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/schemas/range-large.schema.json
new file mode 100644
index 000000000..69b476237
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-large/environment/.deltawire/schemas/range-large.schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 500,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-large/environment/Dockerfile b/labs/20-deltawire/eval/tasks/range-large/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/range-large/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/range-large/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/range-large/environment/task-contract.json b/labs/20-deltawire/eval/tasks/range-large/environment/task-contract.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-large/environment/task-contract.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-large/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/range-large/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-large/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/range-large/instruction.md b/labs/20-deltawire/eval/tasks/range-large/instruction.md
index 36586cf78..9181f4398 100644
--- a/labs/20-deltawire/eval/tasks/range-large/instruction.md
+++ b/labs/20-deltawire/eval/tasks/range-large/instruction.md
@@ -1 +1,28 @@
-# Instruction
+Generate the exact `range-large` dataset defined by public `/task-contract.json`. Write exactly 500 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/range-large/task-spec.json b/labs/20-deltawire/eval/tasks/range-large/task-spec.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-large/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-large/task.toml b/labs/20-deltawire/eval/tasks/range-large/task.toml
index 109209484..4ff5259e3 100644
--- a/labs/20-deltawire/eval/tasks/range-large/task.toml
+++ b/labs/20-deltawire/eval/tasks/range-large/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/range-large.dw.json", ".deltawire/schemas/range-large.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/range-large"
diff --git a/labs/20-deltawire/eval/tasks/range-large/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/range-large/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-large/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/range-large/tests/task-spec.json b/labs/20-deltawire/eval/tasks/range-large/tests/task-spec.json
new file mode 100644
index 000000000..281ba55ac
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-large/tests/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-large.schema.json",
+    "sha256": "2fe7bdc8de996a09ccc651f42611dc88faee60c7390b36c62443ed3e523800b5"
+  },
+  "deltawire_applicability": "expected",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 500,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "range-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-large/tests/test.sh b/labs/20-deltawire/eval/tasks/range-large/tests/test.sh
old mode 100644
new mode 100755
index 06bd98656..9b56450e6
--- a/labs/20-deltawire/eval/tasks/range-large/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/range-large/tests/test.sh
@@ -1,2 +1,7 @@
 #!/bin/bash
-exit 0
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/range-medium/authoritative-schema.json b/labs/20-deltawire/eval/tasks/range-medium/authoritative-schema.json
new file mode 100644
index 000000000..4c0ec1948
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-medium/authoritative-schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 50,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-medium",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-medium/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/range-medium/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-medium/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-medium/environment/.deltawire/schemas/range-medium.schema.json b/labs/20-deltawire/eval/tasks/range-medium/environment/.deltawire/schemas/range-medium.schema.json
new file mode 100644
index 000000000..4c0ec1948
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-medium/environment/.deltawire/schemas/range-medium.schema.json
@@ -0,0 +1,16 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "index": {
+      "maximum": 50,
+      "minimum": 1,
+      "type": "integer"
+    }
+  },
+  "required": [
+    "index"
+  ],
+  "title": "range-medium",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-medium/environment/Dockerfile b/labs/20-deltawire/eval/tasks/range-medium/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/range-medium/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/range-medium/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/range-medium/environment/task-contract.json b/labs/20-deltawire/eval/tasks/range-medium/environment/task-contract.json
new file mode 100644
index 000000000..1c54d3a57
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-medium/environment/task-contract.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-medium.schema.json",
+    "sha256": "c41845095a8893b83b5d3bdc1fd7895b22502932df0590aebce2667002cb38cf"
+  },
+  "deltawire_applicability": "optional",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 50,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "range-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-medium/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/range-medium/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-medium/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/range-medium/instruction.md b/labs/20-deltawire/eval/tasks/range-medium/instruction.md
index 36586cf78..b949f6b4d 100644
--- a/labs/20-deltawire/eval/tasks/range-medium/instruction.md
+++ b/labs/20-deltawire/eval/tasks/range-medium/instruction.md
@@ -1 +1,28 @@
-# Instruction
+Generate the exact `range-medium` dataset defined by public `/task-contract.json`. Write exactly 50 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-medium.schema.json",
+    "sha256": "c41845095a8893b83b5d3bdc1fd7895b22502932df0590aebce2667002cb38cf"
+  },
+  "deltawire_applicability": "optional",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 50,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "range-medium"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/range-medium/task-spec.json b/labs/20-deltawire/eval/tasks/range-medium/task-spec.json
new file mode 100644
index 000000000..1c54d3a57
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-medium/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-medium.schema.json",
+    "sha256": "c41845095a8893b83b5d3bdc1fd7895b22502932df0590aebce2667002cb38cf"
+  },
+  "deltawire_applicability": "optional",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 50,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "range-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-medium/task.toml b/labs/20-deltawire/eval/tasks/range-medium/task.toml
index 73944c1bb..6683c57ef 100644
--- a/labs/20-deltawire/eval/tasks/range-medium/task.toml
+++ b/labs/20-deltawire/eval/tasks/range-medium/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/range-medium.dw.json", ".deltawire/schemas/range-medium.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/range-medium"
diff --git a/labs/20-deltawire/eval/tasks/range-medium/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/range-medium/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-medium/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/range-medium/tests/task-spec.json b/labs/20-deltawire/eval/tasks/range-medium/tests/task-spec.json
new file mode 100644
index 000000000..1c54d3a57
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/range-medium/tests/task-spec.json
@@ -0,0 +1,24 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/range-medium.schema.json",
+    "sha256": "c41845095a8893b83b5d3bdc1fd7895b22502932df0590aebce2667002cb38cf"
+  },
+  "deltawire_applicability": "optional",
+  "family": "range",
+  "generation": {
+    "field": "index",
+    "field_order": [
+      "index"
+    ],
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 50,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "range-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/range-medium/tests/test.sh b/labs/20-deltawire/eval/tasks/range-medium/tests/test.sh
old mode 100644
new mode 100755
index 06bd98656..9b56450e6
--- a/labs/20-deltawire/eval/tasks/range-medium/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/range-medium/tests/test.sh
@@ -1,2 +1,7 @@
 #!/bin/bash
-exit 0
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/rows-large/authoritative-schema.json b/labs/20-deltawire/eval/tasks/rows-large/authoritative-schema.json
new file mode 100644
index 000000000..8f6eb2f59
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-large/authoritative-schema.json
@@ -0,0 +1,25 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "active": {
+      "const": true
+    },
+    "id": {
+      "maximum": 100,
+      "minimum": 1,
+      "type": "integer"
+    },
+    "name": {
+      "pattern": "^User [0-9]+$",
+      "type": "string"
+    }
+  },
+  "required": [
+    "id",
+    "name",
+    "active"
+  ],
+  "title": "rows-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-large/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/rows-large/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-large/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-large/environment/.deltawire/schemas/rows-large.schema.json b/labs/20-deltawire/eval/tasks/rows-large/environment/.deltawire/schemas/rows-large.schema.json
new file mode 100644
index 000000000..8f6eb2f59
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-large/environment/.deltawire/schemas/rows-large.schema.json
@@ -0,0 +1,25 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "active": {
+      "const": true
+    },
+    "id": {
+      "maximum": 100,
+      "minimum": 1,
+      "type": "integer"
+    },
+    "name": {
+      "pattern": "^User [0-9]+$",
+      "type": "string"
+    }
+  },
+  "required": [
+    "id",
+    "name",
+    "active"
+  ],
+  "title": "rows-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-large/environment/Dockerfile b/labs/20-deltawire/eval/tasks/rows-large/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/rows-large/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/rows-large/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/rows-large/environment/task-contract.json b/labs/20-deltawire/eval/tasks/rows-large/environment/task-contract.json
new file mode 100644
index 000000000..aa8f269a8
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-large/environment/task-contract.json
@@ -0,0 +1,29 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/rows-large.schema.json",
+    "sha256": "4a78b0a31f700a97bc98356f53f8a79c417756e3c63fa46c705f4b41986f1961"
+  },
+  "deltawire_applicability": "expected",
+  "family": "rows",
+  "generation": {
+    "constants": {
+      "active": true
+    },
+    "field_order": [
+      "id",
+      "name",
+      "active"
+    ],
+    "name_template": "User {id}",
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 100,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "rows-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-large/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/rows-large/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-large/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/rows-large/instruction.md b/labs/20-deltawire/eval/tasks/rows-large/instruction.md
index 36586cf78..1cb5f3bbe 100644
--- a/labs/20-deltawire/eval/tasks/rows-large/instruction.md
+++ b/labs/20-deltawire/eval/tasks/rows-large/instruction.md
@@ -1 +1,33 @@
-# Instruction
+Generate the exact `rows-large` dataset defined by public `/task-contract.json`. Write exactly 100 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/rows-large.schema.json",
+    "sha256": "4a78b0a31f700a97bc98356f53f8a79c417756e3c63fa46c705f4b41986f1961"
+  },
+  "deltawire_applicability": "expected",
+  "family": "rows",
+  "generation": {
+    "constants": {
+      "active": true
+    },
+    "field_order": [
+      "id",
+      "name",
+      "active"
+    ],
+    "name_template": "User {id}",
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 100,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "rows-large"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/rows-large/task-spec.json b/labs/20-deltawire/eval/tasks/rows-large/task-spec.json
new file mode 100644
index 000000000..aa8f269a8
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-large/task-spec.json
@@ -0,0 +1,29 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/rows-large.schema.json",
+    "sha256": "4a78b0a31f700a97bc98356f53f8a79c417756e3c63fa46c705f4b41986f1961"
+  },
+  "deltawire_applicability": "expected",
+  "family": "rows",
+  "generation": {
+    "constants": {
+      "active": true
+    },
+    "field_order": [
+      "id",
+      "name",
+      "active"
+    ],
+    "name_template": "User {id}",
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 100,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "rows-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-large/task.toml b/labs/20-deltawire/eval/tasks/rows-large/task.toml
index d43ce2486..e44fa5df2 100644
--- a/labs/20-deltawire/eval/tasks/rows-large/task.toml
+++ b/labs/20-deltawire/eval/tasks/rows-large/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/rows-large.dw.json", ".deltawire/schemas/rows-large.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/rows-large"
diff --git a/labs/20-deltawire/eval/tasks/rows-large/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/rows-large/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-large/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/rows-large/tests/task-spec.json b/labs/20-deltawire/eval/tasks/rows-large/tests/task-spec.json
new file mode 100644
index 000000000..aa8f269a8
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-large/tests/task-spec.json
@@ -0,0 +1,29 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/rows-large.schema.json",
+    "sha256": "4a78b0a31f700a97bc98356f53f8a79c417756e3c63fa46c705f4b41986f1961"
+  },
+  "deltawire_applicability": "expected",
+  "family": "rows",
+  "generation": {
+    "constants": {
+      "active": true
+    },
+    "field_order": [
+      "id",
+      "name",
+      "active"
+    ],
+    "name_template": "User {id}",
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 100,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "rows-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-large/tests/test.sh b/labs/20-deltawire/eval/tasks/rows-large/tests/test.sh
old mode 100644
new mode 100755
index 06bd98656..9b56450e6
--- a/labs/20-deltawire/eval/tasks/rows-large/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/rows-large/tests/test.sh
@@ -1,2 +1,7 @@
 #!/bin/bash
-exit 0
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/authoritative-schema.json b/labs/20-deltawire/eval/tasks/rows-medium/authoritative-schema.json
new file mode 100644
index 000000000..2f7d5a458
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-medium/authoritative-schema.json
@@ -0,0 +1,25 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "active": {
+      "const": true
+    },
+    "id": {
+      "maximum": 10,
+      "minimum": 1,
+      "type": "integer"
+    },
+    "name": {
+      "pattern": "^User [0-9]+$",
+      "type": "string"
+    }
+  },
+  "required": [
+    "id",
+    "name",
+    "active"
+  ],
+  "title": "rows-medium",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/rows-medium/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-medium/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/environment/.deltawire/schemas/rows-medium.schema.json b/labs/20-deltawire/eval/tasks/rows-medium/environment/.deltawire/schemas/rows-medium.schema.json
new file mode 100644
index 000000000..2f7d5a458
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-medium/environment/.deltawire/schemas/rows-medium.schema.json
@@ -0,0 +1,25 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "active": {
+      "const": true
+    },
+    "id": {
+      "maximum": 10,
+      "minimum": 1,
+      "type": "integer"
+    },
+    "name": {
+      "pattern": "^User [0-9]+$",
+      "type": "string"
+    }
+  },
+  "required": [
+    "id",
+    "name",
+    "active"
+  ],
+  "title": "rows-medium",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/environment/Dockerfile b/labs/20-deltawire/eval/tasks/rows-medium/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/rows-medium/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/rows-medium/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/environment/task-contract.json b/labs/20-deltawire/eval/tasks/rows-medium/environment/task-contract.json
new file mode 100644
index 000000000..37d30d15c
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-medium/environment/task-contract.json
@@ -0,0 +1,29 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/rows-medium.schema.json",
+    "sha256": "394ece8c17f3627940c75c96ff91df3aecc7a80f5b7eaa5e4876a11a7c29f754"
+  },
+  "deltawire_applicability": "optional",
+  "family": "rows",
+  "generation": {
+    "constants": {
+      "active": true
+    },
+    "field_order": [
+      "id",
+      "name",
+      "active"
+    ],
+    "name_template": "User {id}",
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 10,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "rows-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/rows-medium/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-medium/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/instruction.md b/labs/20-deltawire/eval/tasks/rows-medium/instruction.md
index 36586cf78..5d8e3e352 100644
--- a/labs/20-deltawire/eval/tasks/rows-medium/instruction.md
+++ b/labs/20-deltawire/eval/tasks/rows-medium/instruction.md
@@ -1 +1,33 @@
-# Instruction
+Generate the exact `rows-medium` dataset defined by public `/task-contract.json`. Write exactly 10 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/rows-medium.schema.json",
+    "sha256": "394ece8c17f3627940c75c96ff91df3aecc7a80f5b7eaa5e4876a11a7c29f754"
+  },
+  "deltawire_applicability": "optional",
+  "family": "rows",
+  "generation": {
+    "constants": {
+      "active": true
+    },
+    "field_order": [
+      "id",
+      "name",
+      "active"
+    ],
+    "name_template": "User {id}",
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 10,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "rows-medium"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/task-spec.json b/labs/20-deltawire/eval/tasks/rows-medium/task-spec.json
new file mode 100644
index 000000000..37d30d15c
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-medium/task-spec.json
@@ -0,0 +1,29 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/rows-medium.schema.json",
+    "sha256": "394ece8c17f3627940c75c96ff91df3aecc7a80f5b7eaa5e4876a11a7c29f754"
+  },
+  "deltawire_applicability": "optional",
+  "family": "rows",
+  "generation": {
+    "constants": {
+      "active": true
+    },
+    "field_order": [
+      "id",
+      "name",
+      "active"
+    ],
+    "name_template": "User {id}",
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 10,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "rows-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/task.toml b/labs/20-deltawire/eval/tasks/rows-medium/task.toml
index 7c42b263b..2f2fc533f 100644
--- a/labs/20-deltawire/eval/tasks/rows-medium/task.toml
+++ b/labs/20-deltawire/eval/tasks/rows-medium/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/rows-medium.dw.json", ".deltawire/schemas/rows-medium.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/rows-medium"
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/rows-medium/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-medium/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/tests/task-spec.json b/labs/20-deltawire/eval/tasks/rows-medium/tests/task-spec.json
new file mode 100644
index 000000000..37d30d15c
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/rows-medium/tests/task-spec.json
@@ -0,0 +1,29 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/rows-medium.schema.json",
+    "sha256": "394ece8c17f3627940c75c96ff91df3aecc7a80f5b7eaa5e4876a11a7c29f754"
+  },
+  "deltawire_applicability": "optional",
+  "family": "rows",
+  "generation": {
+    "constants": {
+      "active": true
+    },
+    "field_order": [
+      "id",
+      "name",
+      "active"
+    ],
+    "name_template": "User {id}",
+    "start": 1
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 10,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "rows-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/rows-medium/tests/test.sh b/labs/20-deltawire/eval/tasks/rows-medium/tests/test.sh
old mode 100644
new mode 100755
index 06bd98656..9b56450e6
--- a/labs/20-deltawire/eval/tasks/rows-medium/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/rows-medium/tests/test.sh
@@ -1,2 +1,7 @@
 #!/bin/bash
-exit 0
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/variants-large/authoritative-schema.json b/labs/20-deltawire/eval/tasks/variants-large/authoritative-schema.json
new file mode 100644
index 000000000..bb9577833
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-large/authoritative-schema.json
@@ -0,0 +1,74 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "id": {
+      "const": 1
+    },
+    "type": {
+      "const": "A"
+    },
+    "value": {
+      "enum": [
+        100,
+        110,
+        120,
+        130,
+        140,
+        150,
+        160,
+        170,
+        180,
+        190,
+        200,
+        210,
+        220,
+        230,
+        240,
+        250,
+        260,
+        270,
+        280,
+        290,
+        300,
+        310,
+        320,
+        330,
+        340,
+        350,
+        360,
+        370,
+        380,
+        390,
+        400,
+        410,
+        420,
+        430,
+        440,
+        450,
+        460,
+        470,
+        480,
+        490,
+        500,
+        510,
+        520,
+        530,
+        540,
+        550,
+        560,
+        570,
+        580,
+        590,
+        600
+      ]
+    }
+  },
+  "required": [
+    "id",
+    "type",
+    "value"
+  ],
+  "title": "variants-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-large/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/variants-large/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-large/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-large/environment/.deltawire/schemas/variants-large.schema.json b/labs/20-deltawire/eval/tasks/variants-large/environment/.deltawire/schemas/variants-large.schema.json
new file mode 100644
index 000000000..bb9577833
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-large/environment/.deltawire/schemas/variants-large.schema.json
@@ -0,0 +1,74 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "id": {
+      "const": 1
+    },
+    "type": {
+      "const": "A"
+    },
+    "value": {
+      "enum": [
+        100,
+        110,
+        120,
+        130,
+        140,
+        150,
+        160,
+        170,
+        180,
+        190,
+        200,
+        210,
+        220,
+        230,
+        240,
+        250,
+        260,
+        270,
+        280,
+        290,
+        300,
+        310,
+        320,
+        330,
+        340,
+        350,
+        360,
+        370,
+        380,
+        390,
+        400,
+        410,
+        420,
+        430,
+        440,
+        450,
+        460,
+        470,
+        480,
+        490,
+        500,
+        510,
+        520,
+        530,
+        540,
+        550,
+        560,
+        570,
+        580,
+        590,
+        600
+      ]
+    }
+  },
+  "required": [
+    "id",
+    "type",
+    "value"
+  ],
+  "title": "variants-large",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-large/environment/Dockerfile b/labs/20-deltawire/eval/tasks/variants-large/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/variants-large/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/variants-large/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/variants-large/environment/task-contract.json b/labs/20-deltawire/eval/tasks/variants-large/environment/task-contract.json
new file mode 100644
index 000000000..894fc758a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-large/environment/task-contract.json
@@ -0,0 +1,30 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/variants-large.schema.json",
+    "sha256": "23b704bab1c9f40e7ff93febe34f143b9f2ccd012afaffede927d91f7ca9fac3"
+  },
+  "deltawire_applicability": "expected",
+  "family": "variants",
+  "generation": {
+    "base": {
+      "id": 1,
+      "type": "A",
+      "value": 100
+    },
+    "field_order": [
+      "id",
+      "type",
+      "value"
+    ],
+    "value_step": 10
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 51,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "variants-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-large/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/variants-large/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-large/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/variants-large/instruction.md b/labs/20-deltawire/eval/tasks/variants-large/instruction.md
index 36586cf78..1ac619b42 100644
--- a/labs/20-deltawire/eval/tasks/variants-large/instruction.md
+++ b/labs/20-deltawire/eval/tasks/variants-large/instruction.md
@@ -1 +1,34 @@
-# Instruction
+Generate the exact `variants-large` dataset defined by public `/task-contract.json`. Write exactly 51 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/variants-large.schema.json",
+    "sha256": "23b704bab1c9f40e7ff93febe34f143b9f2ccd012afaffede927d91f7ca9fac3"
+  },
+  "deltawire_applicability": "expected",
+  "family": "variants",
+  "generation": {
+    "base": {
+      "id": 1,
+      "type": "A",
+      "value": 100
+    },
+    "field_order": [
+      "id",
+      "type",
+      "value"
+    ],
+    "value_step": 10
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 51,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "variants-large"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/variants-large/task-spec.json b/labs/20-deltawire/eval/tasks/variants-large/task-spec.json
new file mode 100644
index 000000000..894fc758a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-large/task-spec.json
@@ -0,0 +1,30 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/variants-large.schema.json",
+    "sha256": "23b704bab1c9f40e7ff93febe34f143b9f2ccd012afaffede927d91f7ca9fac3"
+  },
+  "deltawire_applicability": "expected",
+  "family": "variants",
+  "generation": {
+    "base": {
+      "id": 1,
+      "type": "A",
+      "value": 100
+    },
+    "field_order": [
+      "id",
+      "type",
+      "value"
+    ],
+    "value_step": 10
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 51,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "variants-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-large/task.toml b/labs/20-deltawire/eval/tasks/variants-large/task.toml
index 3dc9eb095..f1f7e9502 100644
--- a/labs/20-deltawire/eval/tasks/variants-large/task.toml
+++ b/labs/20-deltawire/eval/tasks/variants-large/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/variants-large.dw.json", ".deltawire/schemas/variants-large.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/variants-large"
diff --git a/labs/20-deltawire/eval/tasks/variants-large/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/variants-large/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-large/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/variants-large/tests/task-spec.json b/labs/20-deltawire/eval/tasks/variants-large/tests/task-spec.json
new file mode 100644
index 000000000..894fc758a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-large/tests/task-spec.json
@@ -0,0 +1,30 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/variants-large.schema.json",
+    "sha256": "23b704bab1c9f40e7ff93febe34f143b9f2ccd012afaffede927d91f7ca9fac3"
+  },
+  "deltawire_applicability": "expected",
+  "family": "variants",
+  "generation": {
+    "base": {
+      "id": 1,
+      "type": "A",
+      "value": 100
+    },
+    "field_order": [
+      "id",
+      "type",
+      "value"
+    ],
+    "value_step": 10
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 51,
+  "schema_version": "task-spec.v1",
+  "size": "large",
+  "task": "variants-large"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-large/tests/test.sh b/labs/20-deltawire/eval/tasks/variants-large/tests/test.sh
old mode 100644
new mode 100755
index 06bd98656..9b56450e6
--- a/labs/20-deltawire/eval/tasks/variants-large/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/variants-large/tests/test.sh
@@ -1,2 +1,7 @@
 #!/bin/bash
-exit 0
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/authoritative-schema.json b/labs/20-deltawire/eval/tasks/variants-medium/authoritative-schema.json
new file mode 100644
index 000000000..1d3ca618d
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-medium/authoritative-schema.json
@@ -0,0 +1,29 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "id": {
+      "const": 1
+    },
+    "type": {
+      "const": "A"
+    },
+    "value": {
+      "enum": [
+        100,
+        110,
+        120,
+        130,
+        140,
+        150
+      ]
+    }
+  },
+  "required": [
+    "id",
+    "type",
+    "value"
+  ],
+  "title": "variants-medium",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/environment/.deltawire/config.json b/labs/20-deltawire/eval/tasks/variants-medium/environment/.deltawire/config.json
new file mode 100644
index 000000000..61a05415a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-medium/environment/.deltawire/config.json
@@ -0,0 +1,12 @@
+{
+  "limits": {
+    "max_output_bytes": 104857600,
+    "max_plan_bytes": 1048576,
+    "max_records": 100000,
+    "max_schema_bytes": 1048576
+  },
+  "plans_dir": ".deltawire/plans",
+  "schemas_dir": ".deltawire/schemas",
+  "state_file": ".deltawire/state.json",
+  "version": "deltawire.config.v1"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/environment/.deltawire/schemas/variants-medium.schema.json b/labs/20-deltawire/eval/tasks/variants-medium/environment/.deltawire/schemas/variants-medium.schema.json
new file mode 100644
index 000000000..1d3ca618d
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-medium/environment/.deltawire/schemas/variants-medium.schema.json
@@ -0,0 +1,29 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "additionalProperties": false,
+  "properties": {
+    "id": {
+      "const": 1
+    },
+    "type": {
+      "const": "A"
+    },
+    "value": {
+      "enum": [
+        100,
+        110,
+        120,
+        130,
+        140,
+        150
+      ]
+    }
+  },
+  "required": [
+    "id",
+    "type",
+    "value"
+  ],
+  "title": "variants-medium",
+  "type": "object"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/environment/Dockerfile b/labs/20-deltawire/eval/tasks/variants-medium/environment/Dockerfile
index e9e7b7a6a..126972c7a 100644
--- a/labs/20-deltawire/eval/tasks/variants-medium/environment/Dockerfile
+++ b/labs/20-deltawire/eval/tasks/variants-medium/environment/Dockerfile
@@ -1 +1,8 @@
 FROM ubuntu:22.04
+COPY .generated/deltawire /usr/local/bin/deltawire
+COPY task-contract.json /task-contract.json
+COPY .deltawire /.deltawire
+COPY verify_plan_contract.py /usr/local/bin/verify_plan_contract.py
+RUN chmod 0755 /usr/local/bin/deltawire /usr/local/bin/verify_plan_contract.py && chmod a-w /task-contract.json /.deltawire/config.json /.deltawire/schemas/*.json
+RUN command -v deltawire && deltawire version
+RUN apt-get update && apt-get install -y python3
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/environment/task-contract.json b/labs/20-deltawire/eval/tasks/variants-medium/environment/task-contract.json
new file mode 100644
index 000000000..a4cc723e8
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-medium/environment/task-contract.json
@@ -0,0 +1,30 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/variants-medium.schema.json",
+    "sha256": "3820180824788284aa6fb14e7c35fa79ac3b4380542f3a5c9ca8ad13037a6438"
+  },
+  "deltawire_applicability": "optional",
+  "family": "variants",
+  "generation": {
+    "base": {
+      "id": 1,
+      "type": "A",
+      "value": 100
+    },
+    "field_order": [
+      "id",
+      "type",
+      "value"
+    ],
+    "value_step": 10
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 6,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "variants-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/environment/verify_plan_contract.py b/labs/20-deltawire/eval/tasks/variants-medium/environment/verify_plan_contract.py
new file mode 100644
index 000000000..205d4fc8a
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-medium/environment/verify_plan_contract.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import argparse, hashlib, json, subprocess
+from pathlib import Path
+
+def sha(path): return hashlib.sha256(Path(path).read_bytes()).hexdigest()
+def write(path, value):
+    if path:
+        target=Path(path); target.parent.mkdir(parents=True,exist_ok=True); target.write_text(json.dumps(value,indent=2,sort_keys=True)+"\n")
+
+def main():
+    p=argparse.ArgumentParser(); p.add_argument("--contract",required=True); p.add_argument("--plan",required=True); p.add_argument("--repo",default="."); p.add_argument("--receipt"); p.add_argument("--deltawire",default="deltawire"); a=p.parse_args()
+    repo=Path(a.repo).resolve(); contract_path=Path(a.contract).resolve(); plan=Path(a.plan).resolve(); contract=json.loads(contract_path.read_text())
+    try: plan_arg=str(plan.relative_to(repo))
+    except ValueError: plan_arg=str(plan)
+    result={"schema_version":"plan-contract-receipt.v1","contract_sha256":sha(contract_path),"plan_sha256":sha(plan),"checks":{},"status":"fail"}
+    try:
+        run=subprocess.run([a.deltawire,"inspect","--repo",str(repo),"--format","json",plan_arg],capture_output=True,text=True)
+        result["inspect_exit_code"]=run.returncode
+        if run.returncode: result["error"]=run.stderr.strip() or run.stdout.strip()
+        else:
+            inspected=json.loads(run.stdout); schema=contract["authoritative_schema"]; schema_path=repo/schema["path"]
+            checks={"output_path":inspected.get("output_path")==contract["output"]["path"],"output_format":inspected.get("output_format")==contract["output"]["format"],"projected_records":inspected.get("projected_records")==contract["record_count"],"schema_path":inspected.get("schema_path")==schema["path"],"schema_exists":schema_path.is_file(),"schema_sha256":schema_path.is_file() and sha(schema_path)==schema["sha256"],"exact_count_assertion":str(contract["record_count"]) in json.dumps(inspected.get("assertion_summary",{}))}
+            result.update({"checks":checks,"inspect":inspected,"status":"pass" if all(checks.values()) else "fail"})
+    except (OSError,ValueError,KeyError) as error: result["error"]=str(error)
+    write(a.receipt,result)
+    if result["status"]!="pass": raise SystemExit(1)
+    print(json.dumps(result,sort_keys=True))
+if __name__=="__main__": main()
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/instruction.md b/labs/20-deltawire/eval/tasks/variants-medium/instruction.md
index 36586cf78..e1ffc057e 100644
--- a/labs/20-deltawire/eval/tasks/variants-medium/instruction.md
+++ b/labs/20-deltawire/eval/tasks/variants-medium/instruction.md
@@ -1 +1,34 @@
-# Instruction
+Generate the exact `variants-medium` dataset defined by public `/task-contract.json`. Write exactly 6 JSON records in canonical order as NDJSON to `testdata/generated/output.ndjson`. Every record, field, value, and ordering rule must match this complete embedded contract:
+
+```json
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/variants-medium.schema.json",
+    "sha256": "3820180824788284aa6fb14e7c35fa79ac3b4380542f3a5c9ca8ad13037a6438"
+  },
+  "deltawire_applicability": "optional",
+  "family": "variants",
+  "generation": {
+    "base": {
+      "id": 1,
+      "type": "A",
+      "value": 100
+    },
+    "field_order": [
+      "id",
+      "type",
+      "value"
+    ],
+    "value_step": 10
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 6,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "variants-medium"
+}
+```
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/task-spec.json b/labs/20-deltawire/eval/tasks/variants-medium/task-spec.json
new file mode 100644
index 000000000..a4cc723e8
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-medium/task-spec.json
@@ -0,0 +1,30 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/variants-medium.schema.json",
+    "sha256": "3820180824788284aa6fb14e7c35fa79ac3b4380542f3a5c9ca8ad13037a6438"
+  },
+  "deltawire_applicability": "optional",
+  "family": "variants",
+  "generation": {
+    "base": {
+      "id": 1,
+      "type": "A",
+      "value": 100
+    },
+    "field_order": [
+      "id",
+      "type",
+      "value"
+    ],
+    "value_step": 10
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 6,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "variants-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/task.toml b/labs/20-deltawire/eval/tasks/variants-medium/task.toml
index 85b4f5fba..17784eb43 100644
--- a/labs/20-deltawire/eval/tasks/variants-medium/task.toml
+++ b/labs/20-deltawire/eval/tasks/variants-medium/task.toml
@@ -1,5 +1,5 @@
 schema_version = "1.3"
-artifacts = []
+artifacts = ["testdata/generated/output.ndjson", ".deltawire/config.json", ".deltawire/plans/variants-medium.dw.json", ".deltawire/schemas/variants-medium.schema.json", ".deltawire/plan-contract-receipt.json"]
 
 [task]
 name = "operatorstack/variants-medium"
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/tests/semantic_oracle.py b/labs/20-deltawire/eval/tasks/variants-medium/tests/semantic_oracle.py
new file mode 100644
index 000000000..00cfbfbc6
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-medium/tests/semantic_oracle.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import json
+import sys
+from pathlib import Path
+
+
+def canonical(spec):
+    g, count, family = spec["generation"], spec["record_count"], spec["family"]
+    if family == "matrix":
+        rows = [{}]
+        for key, allowed in g["dimensions"].items():
+            rows = [{**row, key: value} for row in rows for value in allowed]
+        return [{**row, **g["constants"]} for row in rows]
+    if family == "range":
+        return [{g["field"]: i} for i in range(g["start"], g["start"] + count)]
+    if family == "rows":
+        return [{"id": i, "name": g["name_template"].format(id=i), **g["constants"]} for i in range(g["start"], g["start"] + count)]
+    if family == "variants":
+        return [{**g["base"], "value": g["base"]["value"] + i * g["value_step"]} for i in range(count)]
+    if family == "mixed":
+        return ([{"id": i, "type": "std"} for i in range(1, g["standard_count"] + 1)] +
+                [{"id": i, "type": "exc", "error_code": g["error_code"]} for i in range(g["standard_count"] + 1, count + 1)])
+    return [{"id": i, f'{g["dynamic_field_prefix"]}{i}': i} for i in range(g["id_start"], g["id_start"] + count)]
+
+
+def main():
+    if len(sys.argv) != 3:
+        print('{"exact_match": 0}'); return
+    spec_arg, output_arg = Path(sys.argv[1]), Path(sys.argv[2])
+    if not spec_arg.is_file():
+        candidate = Path(__file__).resolve().parents[1] / "tasks" / sys.argv[1] / "task-spec.json"
+        spec_arg = candidate
+    try:
+        spec = json.loads(spec_arg.read_text(encoding="utf-8"))
+        lines = output_arg.read_text(encoding="utf-8").splitlines()
+        if not lines or any(not line.strip() for line in lines):
+            raise ValueError("empty line")
+        actual = [json.loads(line) for line in lines]
+        passed = actual == canonical(spec)
+    except Exception:
+        passed = False
+    print(json.dumps({"exact_match": 1 if passed else 0}))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/tests/task-spec.json b/labs/20-deltawire/eval/tasks/variants-medium/tests/task-spec.json
new file mode 100644
index 000000000..a4cc723e8
--- /dev/null
+++ b/labs/20-deltawire/eval/tasks/variants-medium/tests/task-spec.json
@@ -0,0 +1,30 @@
+{
+  "authoritative_schema": {
+    "path": ".deltawire/schemas/variants-medium.schema.json",
+    "sha256": "3820180824788284aa6fb14e7c35fa79ac3b4380542f3a5c9ca8ad13037a6438"
+  },
+  "deltawire_applicability": "optional",
+  "family": "variants",
+  "generation": {
+    "base": {
+      "id": 1,
+      "type": "A",
+      "value": 100
+    },
+    "field_order": [
+      "id",
+      "type",
+      "value"
+    ],
+    "value_step": 10
+  },
+  "ordering": "canonical_generation_order",
+  "output": {
+    "format": "ndjson",
+    "path": "testdata/generated/output.ndjson"
+  },
+  "record_count": 6,
+  "schema_version": "task-spec.v1",
+  "size": "medium",
+  "task": "variants-medium"
+}
diff --git a/labs/20-deltawire/eval/tasks/variants-medium/tests/test.sh b/labs/20-deltawire/eval/tasks/variants-medium/tests/test.sh
old mode 100644
new mode 100755
index 06bd98656..9b56450e6
--- a/labs/20-deltawire/eval/tasks/variants-medium/tests/test.sh
+++ b/labs/20-deltawire/eval/tasks/variants-medium/tests/test.sh
@@ -1,2 +1,7 @@
 #!/bin/bash
-exit 0
+OUTPUT_FILE="testdata/generated/output.ndjson"
+if [ ! -f "$OUTPUT_FILE" ]; then
+    mkdir -p /logs/verifier && echo '{"exact_match": 0}' > /logs/verifier/reward.json
+    exit 0
+fi
+mkdir -p /logs/verifier && python3 /tests/semantic_oracle.py /tests/task-spec.json "$OUTPUT_FILE" > /logs/verifier/reward.json
diff --git a/labs/20-deltawire/internal/cli/doctor.go b/labs/20-deltawire/internal/cli/doctor.go
index 2733e7389..5404e1aa0 100644
--- a/labs/20-deltawire/internal/cli/doctor.go
+++ b/labs/20-deltawire/internal/cli/doctor.go
@@ -27,9 +27,9 @@ func runDoctor(args []string) error {
 	_ = ctx
 	_ = os.Getenv
 	_ = filepath.Join
-	
+
 	if ctx.Config != nil {
 		fmt.Printf("Config loaded from %s\n", ctx.Config.StateFile)
 	}
 	return nil
-}
\ No newline at end of file
+}
diff --git a/labs/20-deltawire/internal/cli/helpers.go b/labs/20-deltawire/internal/cli/helpers.go
index a6083d062..94061866d 100644
--- a/labs/20-deltawire/internal/cli/helpers.go
+++ b/labs/20-deltawire/internal/cli/helpers.go
@@ -14,6 +14,7 @@ import (
 	"deltawire/internal/plan"
 	"deltawire/internal/schema"
 	"deltawire/internal/store"
+	"deltawire/internal/strictjson"
 )
 
 type ExecutionContext struct {
@@ -232,10 +233,13 @@ func validatePlanFile(ctx *ExecutionContext, planPath string, skipWrite bool) (*
 		if p.Output.Pretty {
 			footer = []byte("\n]\n") // Adjust as needed
 		}
+		outBytes += int64(len(footer))
+		if outBytes > ctx.Config.Limits.MaxOutputBytes {
+			return nil, errors.New(errors.LimitExceeded, "output size exceeds limit")
+		}
 		if tempFile != nil {
 			tempFile.Write(footer)
 		}
-		outBytes += int64(len(footer))
 	}
 
 	for _, c := range p.Assertions.Coverage {
@@ -295,7 +299,7 @@ func getPlanBytesForHash(path string) []byte {
 	// Canonical plan representation hash
 	b, _ := os.ReadFile(path)
 	var v any
-	json.Unmarshal(b, &v)
+	strictjson.Unmarshal(b, &v)
 	canonical, _ := json.Marshal(v)
 	return canonical
 }
@@ -307,6 +311,13 @@ func resolvePointer(m map[string]any, ptr string) (any, error) {
 	parts := strings.Split(ptr[1:], "/")
 	var current any = m
 	for i, p := range parts {
+		for j := 0; j < len(p); j++ {
+			if p[j] == '~' {
+				if j+1 >= len(p) || (p[j+1] != '0' && p[j+1] != '1') {
+					return nil, fmt.Errorf("invalid json pointer escape sequence in %s", p)
+				}
+			}
+		}
 		p = strings.ReplaceAll(p, "~1", "/")
 		p = strings.ReplaceAll(p, "~0", "~")
 
diff --git a/labs/20-deltawire/internal/cli/render.go b/labs/20-deltawire/internal/cli/render.go
index b70f53f1b..503a2e4fa 100644
--- a/labs/20-deltawire/internal/cli/render.go
+++ b/labs/20-deltawire/internal/cli/render.go
@@ -8,6 +8,7 @@ import (
 	"fmt"
 	"os"
 	"path/filepath"
+	"sort"
 )
 
 func runRender(args []string) error {
@@ -113,24 +114,111 @@ func runRender(args []string) error {
 		return nil
 	}
 
+	type backup struct {
+		originalPath string
+		backupPath   string
+	}
+	var backups []backup
+	var committed []string
+	rollback := func() {
+		for _, path := range committed {
+			_ = os.Remove(path)
+		}
+		for i := len(backups) - 1; i >= 0; i-- {
+			_ = os.Rename(backups[i].backupPath, backups[i].originalPath)
+		}
+	}
+
 	for _, res := range results {
-		outPath, _ := ctx.Store.SafePath(res.OutputPath)
-		os.MkdirAll(filepath.Dir(outPath), 0755)
+		outPath, err := ctx.Store.SafePath(res.OutputPath)
+		if err != nil {
+			rollback()
+			return err
+		}
+		if err := os.MkdirAll(filepath.Dir(outPath), 0755); err != nil {
+			rollback()
+			return errors.New(errors.Internal, "failed to create output directory for %s: %v", outPath, err)
+		}
+
+		if _, err := os.Stat(outPath); err == nil {
+			backupFile, createErr := os.CreateTemp(filepath.Dir(outPath), ".deltawire-backup-*")
+			if createErr != nil {
+				rollback()
+				return errors.New(errors.Internal, "failed to reserve backup for %s: %v", outPath, createErr)
+			}
+			bak := backupFile.Name()
+			if closeErr := backupFile.Close(); closeErr != nil {
+				_ = os.Remove(bak)
+				rollback()
+				return errors.New(errors.Internal, "failed to close backup for %s: %v", outPath, closeErr)
+			}
+			_ = os.Remove(bak)
+			if renameErr := os.Rename(outPath, bak); renameErr != nil {
+				rollback()
+				return errors.New(errors.Internal, "failed to back up %s: %v", outPath, renameErr)
+			}
+			backups = append(backups, backup{outPath, bak})
+		} else if !os.IsNotExist(err) {
+			rollback()
+			return errors.New(errors.Internal, "failed to inspect %s: %v", outPath, err)
+		}
+
 		if err := os.Rename(res.TempOutputPath, outPath); err != nil {
+			rollback()
 			return errors.New(errors.Internal, "atomic write failed for %s: %v", outPath, err)
 		}
+		committed = append(committed, outPath)
 		fmt.Printf("Rendered %s\n", outPath)
 	}
 
 	// Sort state entries
-	// ... (We could sort state entries by PlanPath, but leaving it as appended for simplicity unless strictly required)
-	// Plan says: "state entries sorted by plan path"
-	// We should sort them.
+	sort.Slice(state.Entries, func(i, j int) bool {
+		return state.Entries[i].PlanPath < state.Entries[j].PlanPath
+	})
 
 	// Write state
-	statePath, _ := ctx.Store.SafePath(ctx.Config.StateFile)
-	b, _ := json.MarshalIndent(state, "", "  ")
+	statePath, err := ctx.Store.SafePath(ctx.Config.StateFile)
+	if err != nil {
+		rollback()
+		return err
+	}
+	b, err := json.MarshalIndent(state, "", "  ")
+	if err != nil {
+		rollback()
+		return errors.New(errors.Internal, "failed to encode state: %v", err)
+	}
 	b = append(b, '\n')
-	os.WriteFile(statePath, b, 0644)
+
+	if err := os.MkdirAll(filepath.Dir(statePath), 0755); err != nil {
+		rollback()
+		return errors.New(errors.Internal, "failed to create state directory: %v", err)
+	}
+	tmpState, err := os.CreateTemp(filepath.Dir(statePath), "state-*.tmp")
+	if err != nil {
+		rollback()
+		return errors.New(errors.Internal, "failed to create state temp file: %v", err)
+	}
+	tmpStatePath := tmpState.Name()
+	if _, err := tmpState.Write(b); err != nil {
+		_ = tmpState.Close()
+		_ = os.Remove(tmpStatePath)
+		rollback()
+		return errors.New(errors.Internal, "failed to write state temp file: %v", err)
+	}
+	if err := tmpState.Close(); err != nil {
+		_ = os.Remove(tmpStatePath)
+		rollback()
+		return errors.New(errors.Internal, "failed to close state temp file: %v", err)
+	}
+
+	if err := os.Rename(tmpStatePath, statePath); err != nil {
+		_ = os.Remove(tmpStatePath)
+		rollback()
+		return errors.New(errors.Internal, "failed to rename state file: %v", err)
+	}
+	for _, bk := range backups {
+		_ = os.Remove(bk.backupPath)
+	}
+
 	return nil
 }
diff --git a/labs/20-deltawire/internal/engine/engine.go b/labs/20-deltawire/internal/engine/engine.go
index 73dcdbf52..2be4de18b 100644
--- a/labs/20-deltawire/internal/engine/engine.go
+++ b/labs/20-deltawire/internal/engine/engine.go
@@ -6,6 +6,7 @@ import (
 	"deltawire/internal/plan"
 	"deltawire/internal/schema"
 	"fmt"
+	"regexp"
 )
 
 type RecordCallback func(record map[string]any, genName string, genIndex int) error
@@ -23,6 +24,27 @@ func New(cfg *config.Config, validator *schema.Validator) *Engine {
 	return &Engine{cfg: cfg, validator: validator}
 }
 
+func addSafe(a, b int64) (int64, error) {
+	c := a + b
+	if (c > a) == (b > 0) {
+		return c, nil
+	}
+	return 0, errors.New(errors.PlanInvalid, "integer overflow in count")
+}
+
+func multiplySafe(a, b int64) (int64, error) {
+	if a == 0 || b == 0 {
+		return 0, nil
+	}
+	c := a * b
+	if c/a == b {
+		return c, nil
+	}
+	return 0, errors.New(errors.PlanInvalid, "integer overflow in count")
+}
+
+var dimNameRegex = regexp.MustCompile(`^[a-zA-Z0-9_]+$`)
+
 func (e *Engine) Run(p *plan.Plan, cb RecordCallback) (*Statistics, error) {
 	stats := &Statistics{}
 
@@ -38,7 +60,10 @@ func (e *Engine) Run(p *plan.Plan, cb RecordCallback) (*Statistics, error) {
 		if err != nil {
 			return nil, err
 		}
-		projected += c
+		projected, err = addSafe(projected, c)
+		if err != nil {
+			return nil, err
+		}
 	}
 
 	if projected != p.Assertions.Count {
@@ -83,15 +108,39 @@ func countRecords(g plan.Generator, p *plan.Plan) (int64, error) {
 	switch g.Kind {
 	case "matrix":
 		var total int64 = 1
+		seenDims := make(map[string]bool)
 		for _, d := range g.Dimensions {
+			if !dimNameRegex.MatchString(d.Name) {
+				return 0, errors.New(errors.PlanInvalid, "invalid dimension name: %s", d.Name)
+			}
+			if seenDims[d.Name] {
+				return 0, errors.New(errors.PlanInvalid, "duplicate dimension name: %s", d.Name)
+			}
+			seenDims[d.Name] = true
+
+			opts := 0
+			if len(d.Values) > 0 {
+				opts++
+			}
+			if d.Set != "" {
+				opts++
+			}
+			if d.Range != nil {
+				opts++
+			}
+			if opts != 1 {
+				return 0, errors.New(errors.PlanInvalid, "dimension must have exactly one of values, set, or range")
+			}
+
+			var err error
 			if len(d.Values) > 0 {
-				total *= int64(len(d.Values))
+				total, err = multiplySafe(total, int64(len(d.Values)))
 			} else if d.Set != "" {
 				s, ok := p.Sets[d.Set]
 				if !ok {
 					return 0, errors.New(errors.PlanInvalid, "unknown set: %s", d.Set)
 				}
-				total *= int64(len(s))
+				total, err = multiplySafe(total, int64(len(s)))
 			} else if d.Range != nil {
 				if d.Range.Step == 0 {
 					return 0, errors.New(errors.PlanInvalid, "range step cannot be zero")
@@ -107,7 +156,10 @@ func countRecords(g plan.Generator, p *plan.Plan) (int64, error) {
 				if count < 0 {
 					count = 0
 				}
-				total *= count
+				total, err = multiplySafe(total, count)
+			}
+			if err != nil {
+				return 0, err
 			}
 		}
 		return total, nil
diff --git a/labs/20-deltawire/internal/engine/matrix.go b/labs/20-deltawire/internal/engine/matrix.go
index 9d84874e9..5f50335d7 100644
--- a/labs/20-deltawire/internal/engine/matrix.go
+++ b/labs/20-deltawire/internal/engine/matrix.go
@@ -5,6 +5,7 @@ import (
 	"deltawire/internal/plan"
 	"fmt"
 	"regexp"
+	"strings"
 )
 
 func runMatrix(g plan.Generator, p *plan.Plan, cb func(map[string]any) error) error {
@@ -18,7 +19,8 @@ func runMatrix(g plan.Generator, p *plan.Plan, cb func(map[string]any) error) er
 			dims[i] = p.Sets[d.Set]
 		} else if d.Range != nil {
 			r := d.Range
-			for v := r.Start; ; v += r.Step {
+			v := r.Start
+			for {
 				if r.Step > 0 && v >= r.EndExclusive {
 					break
 				}
@@ -26,14 +28,55 @@ func runMatrix(g plan.Generator, p *plan.Plan, cb func(map[string]any) error) er
 					break
 				}
 				dims[i] = append(dims[i], int64(v))
+
+				// check overflow
+				if r.Step > 0 && v > (9223372036854775807-r.Step) {
+					break
+				}
+				if r.Step < 0 && v < (-9223372036854775808-r.Step) {
+					break
+				}
+				v += r.Step
 			}
 		}
 	}
 
 	state := make([]int, len(dims))
+
+	used := findPlaceholders(g.Record)
+	for _, n := range names {
+		if !used[n] {
+			return errors.New(errors.PlanInvalid, "dimension %s is not referenced in record template", n)
+		}
+	}
+
 	return matrixIter(dims, names, state, 0, g, p, cb)
 }
 
+func findPlaceholders(v any) map[string]bool {
+	used := make(map[string]bool)
+	switch tv := v.(type) {
+	case string:
+		matches := placeholderRegex.FindAllStringSubmatch(tv, -1)
+		for _, m := range matches {
+			used[m[1]] = true
+		}
+	case map[string]any:
+		for _, val := range tv {
+			for k := range findPlaceholders(val) {
+				used[k] = true
+			}
+		}
+	case []any:
+		for _, val := range tv {
+			for k := range findPlaceholders(val) {
+				used[k] = true
+			}
+		}
+	}
+	return used
+}
+
 func matrixIter(dims [][]any, names []string, state []int, depth int, g plan.Generator, p *plan.Plan, cb func(map[string]any) error) error {
 	if depth == len(dims) {
 		env := make(map[string]any)
@@ -95,6 +138,9 @@ func interpolateAny(v any, env map[string]any) (any, error) {
 func interpolateString(s string, env map[string]any) (any, error) {
 	matches := placeholderRegex.FindAllStringSubmatch(s, -1)
 	if len(matches) == 0 {
+		if strings.Contains(s, "${") {
+			return nil, errors.New(errors.PlanInvalid, "unclosed placeholder syntax in %q", s)
+		}
 		return s, nil
 	}
 
@@ -103,7 +149,7 @@ func interpolateString(s string, env map[string]any) (any, error) {
 		key := matches[0][1]
 		val, ok := env[key]
 		if !ok {
-			return nil, fmt.Errorf("unknown placeholder: %s", key)
+			return nil, errors.New(errors.PlanInvalid, "unknown placeholder: %s", key)
 		}
 		return val, nil
 	}
@@ -114,7 +160,12 @@ func interpolateString(s string, env map[string]any) (any, error) {
 		key := m[2 : len(m)-1]
 		val, ok := env[key]
 		if !ok {
-			errOut = fmt.Errorf("unknown placeholder: %s", key)
+			errOut = errors.New(errors.PlanInvalid, "unknown placeholder: %s", key)
+			return m
+		}
+		switch val.(type) {
+		case map[string]any, []any:
+			errOut = errors.New(errors.PlanInvalid, "cannot embed non-scalar value for placeholder: %s", key)
 			return m
 		}
 		return fmt.Sprintf("%v", val)
@@ -122,5 +173,8 @@ func interpolateString(s string, env map[string]any) (any, error) {
 	if errOut != nil {
 		return nil, errOut
 	}
+	if strings.Contains(replaced, "${") {
+		return nil, errors.New(errors.PlanInvalid, "unclosed placeholder syntax in %q", s)
+	}
 	return replaced, nil
 }
diff --git a/labs/20-deltawire/internal/engine/rows.go b/labs/20-deltawire/internal/engine/rows.go
index 8c4954f56..f10a2cc3c 100644
--- a/labs/20-deltawire/internal/engine/rows.go
+++ b/labs/20-deltawire/internal/engine/rows.go
@@ -39,6 +39,13 @@ func setPointer(m *map[string]any, ptr string, val any) error {
 	}
 	parts := strings.Split(ptr[1:], "/")
 	for i, p := range parts {
+		for j := 0; j < len(p); j++ {
+			if p[j] == '~' {
+				if j+1 >= len(p) || (p[j+1] != '0' && p[j+1] != '1') {
+					return fmt.Errorf("invalid json pointer escape sequence in %s", p)
+				}
+			}
+		}
 		p = strings.ReplaceAll(p, "~1", "/")
 		p = strings.ReplaceAll(p, "~0", "~")
 		parts[i] = p
diff --git a/labs/20-deltawire/internal/engine/variants.go b/labs/20-deltawire/internal/engine/variants.go
index 04c633b5a..c95c47e96 100644
--- a/labs/20-deltawire/internal/engine/variants.go
+++ b/labs/20-deltawire/internal/engine/variants.go
@@ -46,6 +46,13 @@ func omitPointer(m *map[string]any, ptr string) error {
 	}
 	parts := strings.Split(ptr[1:], "/")
 	for i, p := range parts {
+		for j := 0; j < len(p); j++ {
+			if p[j] == '~' {
+				if j+1 >= len(p) || (p[j+1] != '0' && p[j+1] != '1') {
+					return fmt.Errorf("invalid json pointer escape sequence in %s", p)
+				}
+			}
+		}
 		p = strings.ReplaceAll(p, "~1", "/")
 		p = strings.ReplaceAll(p, "~0", "~")
 		parts[i] = p
diff --git a/labs/20-deltawire/internal/report/report.go b/labs/20-deltawire/internal/report/report.go
index 2b8e461de..bbf15d09e 100644
--- a/labs/20-deltawire/internal/report/report.go
+++ b/labs/20-deltawire/internal/report/report.go
@@ -3,6 +3,7 @@ package report
 import (
 	"encoding/json"
 	"fmt"
+	"sort"
 	"strings"
 
 	"deltawire/internal/plan"
@@ -58,6 +59,7 @@ func BuildReport(p *plan.Plan, pPath string, pBytes int64, sBytes int64, outByte
 	for k := range primsMap {
 		prims = append(prims, k)
 	}
+	sort.Strings(prims)
 
 	var planAmp, coldAmp string
 	if pBytes > 0 {
diff --git a/labs/20-deltawire/internal/schema/schema.go b/labs/20-deltawire/internal/schema/schema.go
index 7d1cbefd3..1c0fd0eae 100644
--- a/labs/20-deltawire/internal/schema/schema.go
+++ b/labs/20-deltawire/internal/schema/schema.go
@@ -15,12 +15,17 @@ type Validator struct {
 
 // Compile compiles a JSON Schema from a local file, preventing network access.
 func Compile(schemaBytes []byte) (*Validator, error) {
+	parsed := bytesToInterface(schemaBytes)
+	if err := checkNoRemoteRefs(parsed); err != nil {
+		return nil, err
+	}
+
 	c := jsonschema.NewCompiler()
 	// Prevent remote references by blocking non-local loaders and schemas
 	// v6 jsonschema uses Loaders.
 	// We just provide the file in memory.
 	url := "schema.json"
-	if err := c.AddResource(url, bytesToInterface(schemaBytes)); err != nil {
+	if err := c.AddResource(url, parsed); err != nil {
 		return nil, errors.New(errors.SchemaInvalid, "invalid schema resource: %v", err)
 	}
 	sch, err := c.Compile(url)
@@ -33,6 +38,29 @@ func Compile(schemaBytes []byte) (*Validator, error) {
 	return &Validator{sch: sch}, nil
 }
 
+func checkNoRemoteRefs(v any) error {
+	switch v := v.(type) {
+	case map[string]any:
+		if ref, ok := v["$ref"].(string); ok {
+			if strings.HasPrefix(ref, "http://") || strings.HasPrefix(ref, "https://") {
+				return errors.New(errors.SchemaInvalid, "remote references are not permitted: %v", ref)
+			}
+		}
+		for _, val := range v {
+			if err := checkNoRemoteRefs(val); err != nil {
+				return err
+			}
+		}
+	case []any:
+		for _, val := range v {
+			if err := checkNoRemoteRefs(val); err != nil {
+				return err
+			}
+		}
+	}
+	return nil
+}
+
 func (v *Validator) Validate(record map[string]any) error {
 	return v.sch.Validate(record)
 }
diff --git a/labs/20-deltawire/internal/store/store.go b/labs/20-deltawire/internal/store/store.go
index 937b9bee1..d6c20a889 100644
--- a/labs/20-deltawire/internal/store/store.go
+++ b/labs/20-deltawire/internal/store/store.go
@@ -41,6 +41,10 @@ func New(repoRoot string, cfg *config.Config) (*Store, error) {
 	if err != nil {
 		return nil, errors.New(errors.Internal, "invalid repo root")
 	}
+	evalRoot, err := filepath.EvalSymlinks(absRoot)
+	if err == nil {
+		absRoot = evalRoot
+	}
 	return &Store{repoRoot: absRoot, cfg: cfg}, nil
 }
 
@@ -54,12 +58,35 @@ func (s *Store) SafePath(relPath string) (string, error) {
 	}
 	fullPath := filepath.Join(s.repoRoot, relPath)
 	cleanPath := filepath.Clean(fullPath)
-	if !strings.HasPrefix(cleanPath, s.repoRoot) {
-		return "", errors.New(errors.PathEscape, "path traversal detected: %s", relPath)
+
+	resolvedPath := cleanPath
+	for {
+		evalPath, evalErr := filepath.EvalSymlinks(resolvedPath)
+		if evalErr == nil {
+			if resolvedPath == cleanPath {
+				cleanPath = evalPath
+			} else {
+				suffix, relErr := filepath.Rel(resolvedPath, cleanPath)
+				if relErr != nil {
+					return "", errors.New(errors.PathEscape, "could not resolve output path: %s", relPath)
+				}
+				cleanPath = filepath.Join(evalPath, suffix)
+			}
+			break
+		}
+		parent := filepath.Dir(resolvedPath)
+		if parent == resolvedPath {
+			break
+		}
+		resolvedPath = parent
+	}
+
+	rel, err := filepath.Rel(s.repoRoot, cleanPath)
+	if err != nil || rel == ".." || strings.HasPrefix(rel, ".."+string(filepath.Separator)) || filepath.IsAbs(rel) {
+		return "", errors.New(errors.PathEscape, "path traversal detected: %s (clean: %s, root: %s, rel: %s, err: %v)", relPath, cleanPath, s.repoRoot, rel, err)
 	}
 
 	// Split parts to check for .git and .deltawire datasets
-	rel, _ := filepath.Rel(s.repoRoot, cleanPath)
 	parts := strings.Split(rel, string(filepath.Separator))
 	if len(parts) > 0 && parts[0] == ".git" {
 		return "", errors.New(errors.PathEscape, "outputs under .git/ are invalid")
diff --git a/labs/20-deltawire/internal/store/store_test.go b/labs/20-deltawire/internal/store/store_test.go
index ec6657f9a..f1873326d 100644
--- a/labs/20-deltawire/internal/store/store_test.go
+++ b/labs/20-deltawire/internal/store/store_test.go
@@ -1,5 +1,35 @@
 package store
 
-import "testing"
+import (
+	"os"
+	"path/filepath"
+	"testing"
 
-func TestStore(t *testing.T) {}
+	"deltawire/internal/config"
+)
+
+func TestSafePathRejectsSymlinkedParentForNewOutput(t *testing.T) {
+	repo := t.TempDir()
+	outside := t.TempDir()
+	if err := os.Symlink(outside, filepath.Join(repo, "linked")); err != nil {
+		t.Skipf("symlinks unavailable: %v", err)
+	}
+	store, err := New(repo, &config.Config{})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if _, err := store.SafePath("linked/new/output.ndjson"); err == nil {
+		t.Fatal("SafePath accepted a new output beneath a symlink escaping the repository")
+	}
+}
+
+func TestSafePathAllowsDotDotPrefixedName(t *testing.T) {
+	repo := t.TempDir()
+	store, err := New(repo, &config.Config{})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if _, err := store.SafePath("..valid-name"); err != nil {
+		t.Fatalf("SafePath rejected a contained filename beginning with two dots: %v", err)
+	}
+}
diff --git a/labs/20-deltawire/internal/strictjson/strictjson.go b/labs/20-deltawire/internal/strictjson/strictjson.go
index 678a82034..14961ddf2 100644
--- a/labs/20-deltawire/internal/strictjson/strictjson.go
+++ b/labs/20-deltawire/internal/strictjson/strictjson.go
@@ -8,7 +8,7 @@ import (
 )
 
 // Unmarshal strictly parses JSON, rejecting unknown fields, trailing data,
-// and duplicate object keys. It uses json.Number for numbers.
+// and duplicate object keys at every nesting level. It uses json.Number for numbers.
 func Unmarshal(data []byte, v any) error {
 	if err := checkDuplicateKeys(data); err != nil {
 		return err
@@ -19,7 +19,8 @@ func Unmarshal(data []byte, v any) error {
 	if err := dec.Decode(v); err != nil {
 		return err
 	}
-	if dec.More() {
+	var dummy json.RawMessage
+	if err := dec.Decode(&dummy); err != io.EOF {
 		return fmt.Errorf("trailing data found")
 	}
 	return nil
@@ -28,7 +29,16 @@ func Unmarshal(data []byte, v any) error {
 func checkDuplicateKeys(data []byte) error {
 	dec := json.NewDecoder(bytes.NewReader(data))
 	dec.UseNumber()
-	return parseValue(dec)
+	if err := parseValue(dec); err != nil {
+		return err
+	}
+	if _, err := dec.Token(); err != io.EOF {
+		if err == nil {
+			return fmt.Errorf("trailing data found")
+		}
+		return err
+	}
+	return nil
 }
 
 func parseValue(dec *json.Decoder) error {
@@ -39,38 +49,40 @@ func parseValue(dec *json.Decoder) error {
 		}
 		return err
 	}
-	switch v := t.(type) {
-	case json.Delim:
-		if v == '{' {
-			keys := make(map[string]bool)
-			for dec.More() {
-				kt, err := dec.Token()
-				if err != nil {
-					return err
-				}
-				keyStr, ok := kt.(string)
-				if !ok {
-					return fmt.Errorf("expected string key, got %T", kt)
-				}
-				if keys[keyStr] {
-					return fmt.Errorf("duplicate key: %s", keyStr)
-				}
-				keys[keyStr] = true
-				if err := parseValue(dec); err != nil {
-					return err
-				}
+	delim, ok := t.(json.Delim)
+	if !ok {
+		return nil
+	}
+	switch delim {
+	case '{':
+		keys := make(map[string]struct{})
+		for dec.More() {
+			keyToken, err := dec.Token()
+			if err != nil {
+				return err
 			}
-			_, err = dec.Token() // consume '}'
-			return err
-		} else if v == '[' {
-			for dec.More() {
-				if err := parseValue(dec); err != nil {
-					return err
-				}
+			key, ok := keyToken.(string)
+			if !ok {
+				return fmt.Errorf("expected string key, got %T", keyToken)
+			}
+			if _, exists := keys[key]; exists {
+				return fmt.Errorf("duplicate key: %s", key)
+			}
+			keys[key] = struct{}{}
+			if err := parseValue(dec); err != nil {
+				return err
+			}
+		}
+		_, err = dec.Token()
+		return err
+	case '[':
+		for dec.More() {
+			if err := parseValue(dec); err != nil {
+				return err
 			}
-			_, err = dec.Token() // consume ']'
-			return err
 		}
+		_, err = dec.Token()
+		return err
 	}
 	return nil
 }
diff --git a/labs/20-deltawire/internal/strictjson/strictjson_test.go b/labs/20-deltawire/internal/strictjson/strictjson_test.go
new file mode 100644
index 000000000..f51e7496f
--- /dev/null
+++ b/labs/20-deltawire/internal/strictjson/strictjson_test.go
@@ -0,0 +1,26 @@
+package strictjson
+
+import (
+	"strings"
+	"testing"
+)
+
+func TestUnmarshalRejectsDuplicateKeysAtEveryDepth(t *testing.T) {
+	for _, input := range []string{
+		`{"a":1,"a":2}`,
+		`{"outer":{"a":1,"a":2}}`,
+		`{"items":[{"a":1,"a":2}]}`,
+	} {
+		var value any
+		if err := Unmarshal([]byte(input), &value); err == nil || !strings.Contains(err.Error(), "duplicate key: a") {
+			t.Fatalf("Unmarshal(%s) error = %v, want duplicate-key error", input, err)
+		}
+	}
+}
+
+func TestUnmarshalRejectsTrailingData(t *testing.T) {
+	var value any
+	if err := Unmarshal([]byte(`{"a":1} {"b":2}`), &value); err == nil {
+		t.Fatal("Unmarshal accepted trailing JSON value")
+	}
+}
diff --git a/labs/20-deltawire/schemas/deltawire-plan.schema.json b/labs/20-deltawire/schemas/deltawire-plan.schema.json
index f6590d97b..64e4f76ee 100644
--- a/labs/20-deltawire/schemas/deltawire-plan.schema.json
+++ b/labs/20-deltawire/schemas/deltawire-plan.schema.json
@@ -29,7 +29,20 @@
       "type": "array",
       "items": {
         "type": "object",
-        "required": ["kind", "name"]
+        "oneOf": [
+          {
+            "properties": { "kind": { "const": "matrix" } },
+            "required": ["kind", "name"]
+          },
+          {
+            "properties": { "kind": { "const": "rows" } },
+            "required": ["kind", "name"]
+          },
+          {
+            "properties": { "kind": { "const": "variants" } },
+            "required": ["kind", "name"]
+          }
+        ]
       }
     },
     "assertions": {
@@ -55,5 +68,6 @@
       "required": ["count"]
     }
   },
-  "required": ["version", "id", "description", "record_schema", "output", "assertions"]
+  "required": ["version", "id", "description", "record_schema", "output", "assertions"],
+  "additionalProperties": false
 }
\ No newline at end of file
diff --git a/labs/20-deltawire/scripts/check-determinism.sh b/labs/20-deltawire/scripts/check-determinism.sh
index c6598de04..be96fe86b 100755
--- a/labs/20-deltawire/scripts/check-determinism.sh
+++ b/labs/20-deltawire/scripts/check-determinism.sh
@@ -1,5 +1,6 @@
 #!/usr/bin/env bash
 set -euo pipefail
+export GOWORK=off
 
 echo "Checking determinism..."
 
@@ -13,7 +14,6 @@ echo "Testing in $DIR1 and $DIR2"
 
 SRC_DIR="$(git rev-parse --show-toplevel)/labs/20-deltawire"
 cd "$SRC_DIR"
-export GOWORK=off
 go build -o /tmp/deltawire_tmp ./cmd/deltawire
 
 cd "$DIR1"
diff --git a/labs/20-deltawire/scripts/check-production-cleanliness.sh b/labs/20-deltawire/scripts/check-production-cleanliness.sh
new file mode 100755
index 000000000..b7cd64df2
--- /dev/null
+++ b/labs/20-deltawire/scripts/check-production-cleanliness.sh
@@ -0,0 +1,66 @@
+#!/bin/bash
+set -euo pipefail
+
+FAIL=0
+
+check_result_ids() {
+  local root="$1"
+  if [ ! -d "$root" ]; then
+    return 0
+  fi
+  while IFS= read -r -d '' FILE; do
+    if python3 - "$FILE" <<'PY'
+import json
+import sys
+path = sys.argv[1]
+try:
+    data = json.load(open(path, encoding="utf-8"))
+except Exception:
+    sys.exit(0)
+if isinstance(data, dict) and data.get("id") == "mock":
+    print(path)
+    sys.exit(1)
+sys.exit(0)
+PY
+    then
+      :
+    else
+      echo "ERROR: mock Harbor result artifact found in $FILE"
+      FAIL=1
+    fi
+  done < <(find "$root" -type f -name result.json -print0)
+}
+
+check_extracted_trial_ids() {
+  local file="$1"
+  if [ ! -f "$file" ]; then
+    return 0
+  fi
+  if python3 - "$file" <<'PY'
+import json
+import sys
+runs = json.load(open(sys.argv[1], encoding="utf-8"))
+bad = [run.get("pair_id") for run in runs if run.get("trial_id") == "mock"]
+if bad:
+    print(",".join(map(str, bad)))
+    sys.exit(1)
+sys.exit(0)
+PY
+  then
+    :
+  else
+    echo "ERROR: mock trial_id found in $file"
+    FAIL=1
+  fi
+}
+
+check_result_ids "labs/20-deltawire/harbor/jobs"
+check_result_ids "labs/20-deltawire/eval/results/canary-v1"
+check_extracted_trial_ids "labs/20-deltawire/eval/results/canary-v1/extracted_runs.json"
+
+if [ "$FAIL" -eq 1 ]; then
+  echo "Production outputs must not contain mock canary artifacts."
+  exit 1
+fi
+echo "Production cleanliness check passed."
+exit 0
diff --git a/labs/20-deltawire/scripts/check-runtime-boundary.sh b/labs/20-deltawire/scripts/check-runtime-boundary.sh
index 3885eb94a..01e2caabc 100755
--- a/labs/20-deltawire/scripts/check-runtime-boundary.sh
+++ b/labs/20-deltawire/scripts/check-runtime-boundary.sh
@@ -1,5 +1,6 @@
 #!/usr/bin/env bash
 set -euo pipefail
+export GOWORK=off
 
 echo "Checking runtime boundary..."
 
diff --git a/labs/20-deltawire/scripts/validate.sh b/labs/20-deltawire/scripts/validate.sh
index 2dc320535..5081873dd 100755
--- a/labs/20-deltawire/scripts/validate.sh
+++ b/labs/20-deltawire/scripts/validate.sh
@@ -1,5 +1,10 @@
 #!/usr/bin/env bash
 set -euo pipefail
+export GOWORK=off
+cd "$(dirname "$0")/.."
+
+# Cleanup at the end
+trap 'rm -rf .deltawire dist inspect.json inspect.md testdata/generated; git checkout go.mod go.sum 2>/dev/null || true' EXIT
 
 echo "==> gofmt verification"
 if [ -n "$(gofmt -l .)" ]; then
@@ -9,7 +14,10 @@ fi
 
 echo "==> go mod tidy verification"
 go mod tidy
-# skipping git check because we are not committing yet
+if ! git diff --exit-code go.mod go.sum; then
+    echo "go mod tidy changed go.mod or go.sum"
+    exit 1
+fi
 
 echo "==> go test ./..."
 go test ./...
@@ -29,6 +37,9 @@ echo "==> runtime-boundary check"
 echo "==> determinism check"
 ./scripts/check-determinism.sh
 
+echo "==> evaluation contracts and production cleanliness"
+./eval/scripts/validate_dataset.sh
+
 echo "==> build static binary"
 mkdir -p dist
 CGO_ENABLED=0 go build -trimpath -o dist/deltawire ./cmd/deltawire