From 006d3433079621ee8b32be59f2ee610a1409b401 Mon Sep 17 00:00:00 2001 From: hudsonaikins-crown Date: Sat, 3 Oct 2026 20:58:22 -0400 Subject: [PATCH] docs: use AgentShift for ProfitCtl tracking --- .github/pull_request_template.md | 2 +- AGENTS.md | 4 ++-- docs/decisions/PCTL-3-PCTL-14-decision-packet.md | 12 +++++++----- tabellio.validation.json | 2 +- 4 files changed, 11 insertions(+), 9 deletions(-) diff --git a/.github/pull_request_template.md b/.github/pull_request_template.md index a07bac4..79d3267 100644 --- a/.github/pull_request_template.md +++ b/.github/pull_request_template.md @@ -12,7 +12,7 @@ What changed and why. - [ ] Codex review completed and actionable findings addressed - [ ] GitHub checks are passing or failures are explicitly classified -- [ ] Plane issue updated with status, blockers, and validation evidence +- [ ] AgentShift story updated with status, blockers, and validation evidence - [ ] Merge/deploy has explicit human approval ## Release Impact diff --git a/AGENTS.md b/AGENTS.md index 4e2a969..adc0048 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -12,7 +12,7 @@ 2. Lead with findings ordered by severity, then classify each item as `actionable`, `informational`, or `already addressed`. 3. Fix actionable findings first. 4. Re-run affected checks after material fixes. -5. Update Plane with status, blockers, and validation evidence. +5. Update AgentShift with status, blockers, and validation evidence. 6. Use `$check-pr` after local review when the task is PR readiness. ## Review Focus @@ -26,7 +26,7 @@ - Required schema, semantic, workflow, operational, and security evidence must pass. `blocked` is not `passed`; any new commit invalidates old evidence. - Pricing correctness, simulation integrity, maintained reference cost standards, CLI output, structured LLM parsing, and zero-cost validation are explicit boundaries. - Never call OpenRouter, deploy, publish, or mutate billing during validation. Upload generated evidence from CI; do not commit it. -- Track rollout and failures in Plane item `PCTL-17`. +- Track rollout and failures in AgentShift. `PCTL-17` remains the legacy acceptance ID in the validation manifest; verify the corresponding AgentShift story before posting updates. ## Code cleanup diff --git a/docs/decisions/PCTL-3-PCTL-14-decision-packet.md b/docs/decisions/PCTL-3-PCTL-14-decision-packet.md index deefc9e..6c9e037 100644 --- a/docs/decisions/PCTL-3-PCTL-14-decision-packet.md +++ b/docs/decisions/PCTL-3-PCTL-14-decision-packet.md @@ -4,7 +4,9 @@ Status: accepted by Hudson in Codex task `019fb11b-6f69-7d41-aabb-c51588d50989` Prepared: 2026-07-31 -Execution authority: Plane PCTL-3, PCTL-14, and PCTL-2 +Current tracker: AgentShift. The PCTL identifiers below are legacy planning references, not verified AgentShift story IDs. + +Historical execution authority: PCTL-3, PCTL-14, and PCTL-2 Implementation authority: PCTL-2 only; merge, release, deployment, live provider access, spend, and secrets remain excluded @@ -22,7 +24,7 @@ This acceptance clears product and provenance definition. It authorizes PCTL-2 i ## Known Facts - `main` presents ProfitCtl as a local CLI for pricing, unit economics, and recurring-margin simulation. -- Plane PCTL-3 and the proposed portfolio roadmap define a broader control loop: forecast, actual, variance, attribution, and action, with Upstash idle polling as first slice. +- The legacy PCTL-3 record and the proposed portfolio roadmap define a broader control loop: forecast, actual, variance, attribution, and action, with Upstash idle polling as first slice. - Dirty canonical branch `docs/profitctl-oss-company-plan` proposes a different public center: “unit economics as code for AI-native software teams,” agent-decision demos, and a 120-day adoption plan. PCTL-10 and PCTL-16 preserve its work, but no active task, PR, or accepted product decision owns that direction. - Canonical checkout is modified across product, open-core, growth, legal, standards-judge, skill, and website surfaces. It remains read-only. - Repository identity is already `github.com/IntelIP/ProfitCtl`; installed binary name is already `profitctl`. @@ -74,7 +76,7 @@ Keep current public product boundary. Ship stable CLI identity and agent-facing Benefits: smallest product change; closest to committed docs and current runtime. -Tradeoff: invalidates current Definition-to-MVP Plane sequence and leaves “cost intelligence” as planning language rather than an executable control loop. +Tradeoff: invalidates then-current Definition-to-MVP planning sequence and leaves “cost intelligence” as planning language rather than an executable control loop. ## Option C — Agent Cost OS Public Wedge First @@ -175,7 +177,7 @@ If accepted: If amended: record exact changed clauses before implementation. -If rejected: choose Option B or C and re-sequence Plane dependencies before implementation. +If rejected: choose Option B or C and re-sequence dependencies in AgentShift before implementation. ## Acceptance Record @@ -190,4 +192,4 @@ Hudson accepted Option A in full on 2026-07-31: - Checksum-verified GitHub Release installer is primary; Bun/npm is secondary; Go paths are explicit developer fallbacks. - `--version` and `doctor` are required. -Plane PCTL-3 and PCTL-14 record the accepted outcome and evidence contracts. PCTL-2 implementation proceeds on isolated branch `codex/pctl-2-stable-cli-identity`; the dirty canonical company-plan checkout remains separate and read-only. +The legacy PCTL-3 and PCTL-14 records capture the accepted outcome and evidence contracts. PCTL-2 implementation proceeds on isolated branch `codex/pctl-2-stable-cli-identity`; the dirty canonical company-plan checkout remains separate and read-only. diff --git a/tabellio.validation.json b/tabellio.validation.json index f4e9e98..fe49962 100644 --- a/tabellio.validation.json +++ b/tabellio.validation.json @@ -5,7 +5,7 @@ "requireEntireCheckpoint": false, "acceptance": { "id": "PCTL-17", - "source": "plane", + "source": "agentshift", "risk": "critical", "outcomes": [ "Cost configuration and provenance contracts reject malformed or unsupported inputs.",