DA
Decision Autopsy
🔵 Stable🕐 updated 2026-10-06
✅ 4.8/5
🔷 SkillSpec L3
pm-warroom
Judge a past decision by its PROCESS, not its outcome — because good decisions lose and bad decisions win, and teams that can't tell the difference learn the wrong lessons. Use when reviewing a big call after the fact (a bet that failed, a pass that haunts, a hire, a pivot) and the room is about to conclude 'it failed so it was wrong.' Produces a process-forensics report: what was knowable then, the quality grade of the decision as-made, the luck accounting, and the ONE process change worth keeping.
▶ Run it free — no key needed
📝 Grade your existing draft
⬇ Add to Claude (works in Duet)
📱 Use as an app
View SKILL.md ↗
What changed
🗣 Say this to your agent
“Was this a bad decision or just bad luck?”“Review the call we made last year.”“Judge this decision by its process, not the outcome.”“What should we learn from this failed bet?”
What to give it
▸The decision — what was decided, when, by whom, and what the live alternatives were.
▸What was knowable at the time — the information, constraints, and time pressure as of the decision date. Be strict: things learned afterward go in a separate pile, and the autopsy will police the boundary.
▸The outcome — what actually happened, so the luck accounting has something to account.
✅ The bar it holds itself to
Every skill in this library self-verifies — these are this skill's own quality checks, straight from its definition.
✓The decision grade was assigned from the knowability ledger BEFORE outcome discussion, and the report's structure shows it
✓At least one hindsight contamination is caught and named — reviews without any are usually not looking
✓The luck paragraph commits to a rough proportion, with reasoning — "some luck was involved" is evasion
✓The process change is executable next quarter and testable ("did we do it?"), not a value statement
✓If the decision was 🟢 and the outcome bad, the report says the uncomfortable sentence plainly: "do it again"
⚠️ What it refuses to do
Do not let the outcome leak into the grade — a bad result may not appear as evidence of a bad decision anywhere in the report
Do not run an autopsy as a trial — no verdicts on people; the unit of analysis is the process that any competent person was embedded in
Do not conclude "we were unlucky" without the ledger to earn it — luck is the residual after process is examined, never the headline
Do not extract more than one lesson — the second-best lesson dilutes the best one
Do not autopsy decisions younger than their outcome — if the result isn't actually in yet, this is a premortem's job
Install
npx pm-claude-skills add --agent claude # or codex · cursor · gemini · hermes
# or one-line MCP (every skill, any client):
claude mcp add pm-skills -- npx -y -p pm-claude-skills pm-claude-skills-mcpRelated skills
🔌 Embed this skill
Drop this on your blog, docs, or site — it renders a "Run this skill" card:
<div data-pm-skill="decision-autopsy"></div>
<script src="https://mohitagw15856.github.io/pm-claude-skills/embed.js" async></script>💬 Discussion
Decision Autopsy is one of 1326 open-source professional AI agent skills — all SkillSpec L3.
Try them all in the browser · ⭐ Star on GitHub · Browse the full catalog