What problem does it solve?
It helps you capture what reviewers actually caught in a merged PR and convert that signal into concrete improvements, so the same issues don’t repeat in future changes.
Core Features & Use Cases
- PR review-to-lessons analysis: Detects the merged PR, fetches metadata, review comments, and review summaries, then extracts actionable lessons that indicate missing skills, specialists, or checklist coverage.
- Pattern-to-routing diagnosis: Classifies findings into pattern-violation vs correctness-bug tracks, then directs the postmortem into the appropriate diagnosis path (specialist vs logic reviewer).
- Self-improving skillbase: Cross-references existing skills/specialists, edits the relevant skill/specialist files to close gaps, and produces a structured report with patterns found, changes made, and judgment-needed items.
Quick Start
Run the postmortem after a PR merges by asking it to run a postmortem for that PR, e.g., “Run a postmortem for PR 1234.”