fix(workbench): share scan history across linked worktrees

README.md

@@ -49,7 +49,8 @@ container credentials, and the default AWS credential chain.
 
 Local sign-in honors Codex's configured credential backend, including a system
 keyring required by a managed device. Codex Security keeps login and scan
-credentials in the same private, persistent state directory.
+sign-in scoped to a dedicated Codex home in the same private, persistent state
+directory.
 
 If both a ChatGPT sign-in and an API key are available, interactive scans ask
 which credential to use. CI and other noninteractive scans keep the existing
@@ -67,12 +68,26 @@ keys:
 unset OPENAI_API_KEY CODEX_API_KEY
 ```
 
-Scan history is stored in the Codex Security workbench state directory. If that
-directory cannot be written, set `CODEX_SECURITY_STATE_DIR` to a writable
-directory outside the repository.
+Scan history and saved findings are stored in one Codex Security state database.
+Linked Git worktrees are discovered automatically and grouped when they use the
+same state directory. Leave `CODEX_SECURITY_STATE_DIR` unset, or select one
+stable, writable directory outside the repository. Changing or unsetting it
+selects separate history and an isolated Codex credential home and sign-in
+scope; restore the previous value to reopen its existing scans and sign-in.
+
+```bash
+# Run from another linked Git worktree:
+npx @openai/codex-security scans list
+npx @openai/codex-security findings list
+
+# Reopen an existing, separately selected state directory:
+export CODEX_SECURITY_STATE_DIR=/path/to/existing/codex-security-state
+npx @openai/codex-security scans list
+```
 
 `findings list [repository]` shows open findings across a repository's scans
-and identifies findings not confirmed in its latest scan.
+and identifies findings not confirmed in its latest completed scan across
+linked worktrees.
 
 `scans compare BEFORE_SCAN_ID AFTER_SCAN_ID` automatically matches findings by
 root cause, reuses saved matches, and identifies new, persisting, reopened,

sdk/typescript/README.md

@@ -426,7 +426,7 @@ The CLI and SDK recognize the following user-configurable environment:
 | `OPENAI_API_KEY`, `CODEX_API_KEY`                                           | Scan authentication; `OPENAI_API_KEY` wins when both are present.                             |
 | `CODEX_SECURITY_LOG_LEVEL`                                                  | CLI-only; set to `debug` for verbose diagnostics.                                             |
 | `LOG_LEVEL`                                                                 | CLI-only fallback when `CODEX_SECURITY_LOG_LEVEL` is unset.                                   |
-| `CODEX_SECURITY_STATE_DIR`                                                  | Override the private scan-history, workbench, and default artifact directory.                 |
+| `CODEX_SECURITY_STATE_DIR`                                                  | Select the history database, artifact directory, and dedicated Codex credential home.         |
 | `CODEX_HOME`                                                                | Set the ambient Codex home for file-backed sign-in and default state; defaults to `~/.codex`. |
 | `CODEX_CLI_PATH`                                                            | Use another Codex executable for authentication, plugin setup, scans, and nested workers.     |
 | `PYTHON`                                                                    | Select a Python interpreter when `--python` or SDK `pythonPath` is not set.                   |
@@ -538,11 +538,15 @@ same command to resume.
 
 ### Scan history and reruns
 
-`npx @openai/codex-security scans list` lists scans for the current repository. Pass a
-repository path to inspect another checkout, `--scan-root DIR` to list scans
-whose artifacts are under a particular root. `scans show SCAN_ID` includes the
-scan configuration, results, coverage, and artifact locations. Add
-`--show-linked-findings` to include finding links from previous scans.
+`npx @openai/codex-security scans list` lists scans for the current repository.
+Linked Git worktrees are discovered automatically and grouped when they share
+the selected state database. Findings indicate whether they were confirmed in
+the repository's latest completed scan across those linked worktrees. Pass a
+repository path to inspect another checkout. `--scan-root DIR` only filters
+scans already recorded in that database by their artifact directory; it never
+imports scan results from another state directory. `scans show SCAN_ID`
+includes the scan configuration, results, coverage, and artifact locations.
+Add `--show-linked-findings` to include finding links from previous scans.
 
 `scans logs SCAN_ID` shows complete session events from the scan and its
 workers, which can include source code and credentials.
@@ -553,9 +557,27 @@ least eight characters.
 Scan history uses `$CODEX_SECURITY_STATE_DIR/workbench.sqlite3` when
 `CODEX_SECURITY_STATE_DIR` is set. Otherwise, it uses
 `$CODEX_HOME/state/plugins/codex-security/workbench.sqlite3`; `CODEX_HOME`
-defaults to `~/.codex`. Scan credentials are never stored in the scan
-configuration. Recorded failure summaries and bulk-scan receipts omit messages
-that contain recognizable credentials.
+defaults to `~/.codex`. Saved findings use the same selected database.
+Changing or unsetting `CODEX_SECURITY_STATE_DIR` selects separate scan history
+and an isolated Codex credential home and sign-in scope; scans from the
+previous state remain hidden until you select that state again. Keep the
+setting stable across linked worktrees to share scans and findings
+automatically.
+
+```bash
+# Inspect shared history from another linked Git worktree:
+npx @openai/codex-security scans list
+npx @openai/codex-security findings list
+
+# Reopen an existing state directory and its sign-in:
+export CODEX_SECURITY_STATE_DIR=/path/to/existing/codex-security-state
+npx @openai/codex-security scans list
+npx @openai/codex-security findings list
+```
+
+Scan credentials are never stored in the scan configuration. Recorded failure
+summaries and bulk-scan receipts omit messages that contain recognizable
+credentials.
 
 The scan sandbox permits writes to the selected state directory so SQLite can
 maintain its database and journal files. If the host itself cannot write to the

sdk/typescript/_bundled_plugin/scripts/workbench_feedback.py

@@ -16,12 +16,24 @@
     FINDING_SUMMARY_BYTES,
     FINDING_TITLE_BYTES,
 )
+from workbench_native_indexes import _indexed_findings, repository_target_ids
 from workbench_validation import bounded_output_text
 
 
 def get_scan_feedback(connection: sqlite3.Connection, scan: sqlite3.Row) -> dict[str, Any]:
+    target_ids = sorted(repository_target_ids(connection, scan["target_id"]))
+    if not target_ids:
+        return {"scanId": scan["id"], "targetId": scan["target_id"], "falsePositives": []}
+
+    indexed_findings = {
+        finding_id: finding
+        for finding in _indexed_findings(connection)
+        if finding["target_id"] in target_ids
+        for finding_id in finding["matched_finding_ids"]
+    }
+    target_placeholders = ", ".join("?" for _ in target_ids)
     rows = connection.execute(
-        """
+        f"""
         WITH ranked_decisions AS (
             SELECT findings.id AS finding_id, findings.fingerprint, findings.rule_id,
                 findings.identity_anchor, findings.identity_instance, occurrences.title,
@@ -49,7 +61,7 @@ def get_scan_feedback(connection: sqlite3.Connection, scan: sqlite3.Row) -> dict
                     candidate.sort_order
                 LIMIT 1
             )
-            WHERE source_scans.target_id = ?
+            WHERE source_scans.target_id IN ({target_placeholders})
                 AND source_scans.id != ?
                 AND source_scans.status = 'complete'
         )
@@ -61,12 +73,22 @@ def get_scan_feedback(connection: sqlite3.Connection, scan: sqlite3.Row) -> dict
             AND note IS NOT NULL
             AND trim(note) != ''
         ORDER BY updated_at DESC, source_completed_at DESC, source_scan_id DESC, finding_id DESC
-        LIMIT 50
         """,
-        (scan["target_id"], scan["id"]),
+        (*target_ids, scan["id"]),
     )
     false_positives = []
+    reviewed_components: set[str] = set()
     for row in rows:
+        finding = indexed_findings.get(row["finding_id"])
+        if (
+            finding is None
+            or finding["status"] != "closed"
+            or finding["close_reason"] != "false_positive"
+            or finding["occurrence_id"] in reviewed_components
+        ):
+            continue
+        reviewed_components.add(finding["occurrence_id"])
+
         identity = {"anchor": row["identity_anchor"]}
         if row["identity_instance"] is not None:
             identity["instance"] = row["identity_instance"]
@@ -91,6 +113,8 @@ def get_scan_feedback(connection: sqlite3.Connection, scan: sqlite3.Row) -> dict
                 "updatedAt": row["updated_at"],
             }
         )
+        if len(false_positives) == 50:
+            break
     return {"scanId": scan["id"], "targetId": scan["target_id"], "falsePositives": false_positives}