cairn-ai 0.2.1 → 0.2.3

package/templates/PLAN.md

@@ -4,6 +4,7 @@ Short index of work plans. Link detailed plans under `docs/plan/`.
 
 ## Planning Rules
 
+- For implementation or continued execution, write an initial plan with `triage-plan` active, synchronize it to native UI plan/goal tools and the repository goal before exploration, then update it after triage.
 - Detect repository stack and required LSP/check tools before implementation.
 - Install or bootstrap missing required tools before declaring them unavailable.
 - Heavy Path completion requires explicit automated test evidence.

package/templates/work-plan.md

@@ -1,5 +1,11 @@
 # Plan: <topic>
 
+## Plan Phase
+
+- Initial: decision-complete for `triage-plan`; not implementation-ready.
+- Finalized: decision-complete for implementation after planned triage.
+- Codex UI synchronization: `update_plan` status and `create_goal` status, or an explicit unavailable reason.
+
 ## Goal
 
 Describe the observable result.
@@ -74,6 +80,12 @@ Describe the observable result.
 
 ## Execution Guardrails
 
+- Test contract first: requirements, invariants, boundaries, failure modes, expected initial failure, and authoritative tool result.
+- Token budget: high reasoning for plan/triage/test design; bounded minimum implementation; concise tool-derived verification summaries.
+- Verification ladder: focused tests, one final full check, then package dry-run if no files changed. Inspect lifecycle scripts first; run normal `npm pack --dry-run` by default. Content-producing or unknown scripts must never use `--ignore-scripts`; only absent or proven content-neutral scripts may use it while evidence remains fresh.
+- Evidence freshness: relevant changes after verification make affected evidence stale.
+- Verification execution: `goal verify` defaults to 600,000 ms (10 minutes), is capped at 3,600,000 ms (1 hour), and records only when the starting goal/task identity and pre/post watched fingerprint match.
+- Evidence boundary: choose the semantic boundary whose watch set and command prove the claimed behavior.
 - Dry-run or check mode before external-state mutation:
 - No dry-run available, if applicable:
 - Verification loop budget: two passes per task by default.
@@ -81,6 +93,15 @@ Describe the observable result.
 
 ## Module Tasks
 
+### Task 0: planned triage and plan finalization
+
+- Task ID: `triage-plan`
+- Initial status: `active`
+- Contract: execute the investigation scope and Heavy Path criteria declared by the initial plan, then update this file and the Codex UI plan to the decision-complete implementation revision.
+- Required evidence records:
+  - `planArtifact`: command, exit code 0, timestamp, goal ID, task ID, and plan ID.
+  - `triageDecision`: command, exit code 0, timestamp, goal ID, task ID, and plan ID.
+
 ### Task 1: <module>
 
 - Task ID: `task-1`
@@ -93,7 +114,7 @@ Describe the observable result.
 - Dry-run or check command:
 - Module acceptance verification:
 - Surface integration verification:
-- Required receipts:
+- Required evidence records:
   - `moduleAcceptance`: command, exit code 0, timestamp, goal ID, task ID, and plan ID.
   - `surfaceIntegration`: command or QA artifact, exit code 0, timestamp, goal ID, task ID, and plan ID.
 - Assigned agent ID, if delegated:
@@ -106,7 +127,7 @@ Describe the observable result.
 - Tests:
 - Module acceptance:
 - Surface integration:
-- Goal final review receipt:
+- Goal final review evidence record:
 - Verification pass count:
 - Blocker after two failed passes: