sdd-distill-lessons

You can automate lesson extraction by reading spec retrospective content to identify and deduplicate potential lessons, then requesting user confirmation before adding them to a specs/lessons.md document for team knowledge sharing.

Lesson deduplication works by grouping extracted retrospective learnings based on semantic similarity, ensuring only unique insights are presented for user review before promotion to your final documentation.

The best way to promote lessons is to use an automated skill that suggests select retrospective learnings for promotion to specs/tech-stack.md, performing an immediate code review upon adding them to ensure accuracy.

Yes, user input is required for decision-making. The distillation process requests user confirmation to select which deduplicated lessons should be included in the specs/lessons.md document.

Yes, automated code review is triggered immediately upon adding new lessons to your lessons file, ensuring that newly captured spec-driven development learnings are validated right away.

You should not use automated lessons distillation if your team lacks existing spec retrospective content or does not operate within a spec-driven development framework, as it requires context files and user input.