post-mortem
Post-Mortem Skill
Purpose: Wrap up completed work — validate it shipped correctly, extract learnings, process the knowledge backlog, activate high-value insights, and retire stale knowledge.
Runtime note: Hook-driven closeout is runtime-dependent. Claude/OpenCode can wire Phase 2-5 maintenance through lifecycle hooks. Codex CLI v0.115.0+ supports native hooks (same behavior). For older Codex versions without hook surfaces, finish closeout with
ao codex stop.
Loop position
Move 7 (capture evidence + learning, then ratchet) of the operating loop. Two outputs per loop turn: evidence (test names, snapshot keys, council verdicts, citation events) recorded against the bead and .agents/flywheel/; learnings promoted only under the ratchet rules — noticed once stays in the handoff, repeats twice goes to .agents/learnings/, changes future behavior updates a SKILL.md or template, must-never-regress becomes a gate, core doctrine promotes into PRODUCT.md/GOALS.md/docs/cdlc.md. Most observations die at handoff. That is correct.
Route a promoted learning to the WEAKEST enforcement surface that actually changes behavior — this is the ladder, by strength, not three copies of the same note:
| Surface | Strength | Promote here when |
|---|---|---|
AGENTS.md/CLAUDE.md |
always-on context | doctrine relevant to most turns in this repo |
| a SKILL.md | JIT, model-invoked | contextual judgment that fires on a trigger |
| a gate/hook | mechanical, un-skippable | must-never-regress; cannot be left to judgment |