gsd-extract-learnings

To extract learnings from phase artifacts, you parse completed PLAN, SUMMARY, VERIFICATION, UAT, and STATE files to consolidate decisions, lessons, and patterns into a single LEARNINGS.md knowledge file.

Consolidating retrospective decisions and rationale into a markdown file requires reading completed phase artifacts and extracting key tradeoffs, unexpected outcomes, and recurring themes into a structured LEARNINGS.md output.

Yes, generating a consolidated learnings markdown file requires completed execution phase artifacts, specifically PLAN.md, SUMMARY.md, VERIFICATION.md, UAT.md, and STATE.md files to parse for structured learning capture.

You can capture unexpected outcomes and recurring patterns by mining completed phase artifacts for surprises and lessons, which are then structured into a reusable markdown knowledge artifact for future cycles.

You should generate a LEARNINGS.md file when a project execution phase is complete and you need to consolidate phase artifacts like SUMMARY and VERIFICATION files to prevent repeating mistakes in the next cycle.