Capture Engineering Learnings

Document engineering learnings by reviewing git diff and log, then drafting routed learning entries for approval to prevent repeating non-obvious pitfalls. This captures outcomes as clear, reusable knowledge for future implementation.

Deduplicate engineering knowledge entries by checking existing Learnings/ and CLAUDE.md files before writing. This prevents overlapping entries and maintains a clean, reusable knowledge base for the team.

Trace issue resolution links in API integration documentation by cross-referencing tracked issues only when an issue-triggered learning is confidently identified. This keeps traceability optional, linking 'Resolves' entries to issues accurately.

Save non-obvious platform learnings into the Learnings/ directory, and update reusable team conventions in CLAUDE.md. Record future work in Plans/, routing knowledge to the right place for team access.

Yes, you can capture debugging and refactoring learnings for WASM behavior. The process reviews recent development work, drafts proposed learning items for approval, and deduplicates against existing notes to accelerate future implementation.

Avoid automated engineering learning capture when tasks lack recent git diff or log data, as the process requires reviewing these to draft proposed learning items. It is also unsuited for tasks lacking non-obvious pitfalls or reusable conventions.