retrospective-codify

To codify task insights into reusable rules, you pair initial failures with final solutions and generate artifacts like ast-grep rules or CLAUDE.md updates. This stabilizes learning and prevents repeating pitfalls across projects.

The best way to prevent repeated mistakes is extracting lessons from completed tasks and converting them into reusable artifacts. This process applies classification checks to avoid duplication and outputs structured rules for future automation.

You generate ast-grep rules by documenting the bridging insight between a failure and its working solution. The Skill uses prescribed templates to produce ready-to-use automation artifacts from these paired task outcomes.

Yes, this retrospective workflow enforces a dedup check to classify insights. It evaluates existing rules or skills to decide whether to create a new artifact or append to an existing one, preventing redundant outputs.

You need to update CLAUDE.md with retrospective artifacts when a dedup check determines an existing entry should be appended. This captures knowledge from complex tasks to inform future onboarding and automation.