kb-postmortemlisted

Post-publish retrospective on one ingested source — collect deterministic proxy metrics (ingest-stats), refresh the failure-signal backlog with increment context (snapshot the old backlog.yaml BEFORE skill-mine, then diff), read the digest routing deviations, and write one standardized report with recommendations only (skill-evolve triggers, gate/truth updates, proposals-resolve dry-run commands). Use when the user says "postmortem this ingest / 复盘这次入库 / how did this book's ingest go / ingest retrospective for <source>". Only for a source already published; "run a KB QA / audit coverage" is kb-qa, "semantic health check" is wiki-lint-semantic, and "distill this failure into the skill" is skill-evolve — none of those trigger this.
Iabstergo1/pdf-to-study-kb · ★ 1 · AI & Automation · score 74

Install: claude install-skill Iabstergo1/pdf-to-study-kb

# kb-postmortem — post-publish ingest retrospective (report-only) After a source publishes, turn the manual "lessons learned" pass into a standard step: deterministic proxy metrics + digest deviations + backlog delta → one report. **Recommendations only — every follow-up action (skill-evolve, gate edits, `proposals-resolve --apply`) is decided by the human.** Project truth: `CLAUDE.md`. ## 1. Triggers / Non-triggers - **Triggers:** "postmortem <source>", "复盘这次入库", "how did this ingest go", "ingest retrospective / review this book's ingest". - **Non-triggers (never fire):** - "run a KB QA / audit coverage / spot-check evidence / run the Q-chain" → `kb-qa`. - "semantic health check / contradictions / L4 / Q2 added value" → `wiki-lint-semantic`. - "distill this failure into the skill / evolve a skill / handle backlog item N" → `skill-evolve` (this skill only **recommends** it). - "work the review queue item by item" → `kb-review`. - the source is not published yet → finish the ingest (lint gate) first; this skill never runs mid-ingest. ## 2. Inputs - `source_id` — must show `lint/published` in `status`. - Proxy metrics via `ingest-stats` (windows, stage durations and reruns, lint failures ≈ rollbacks, pages estimate, violations by kind). Token/cost numbers are not available — never invent them. - `pipeline-workspace/staging/<src>/digest.md`: the per-chapter content-routing table + logged deviations. - The old `pipeline-workspace/skill-evolution/backlog.yaml` — **r