architecture-refinementlisted

# Architecture Refinement Skill > Meta-skill for maintaining and evolving the AI assistant's cognitive architecture through deliberate documentation and pattern extraction. ## Purpose This skill enables the AI assistant to: - Recognize when a learning session produces architecture-worthy insights - Document patterns in the appropriate .github/ location (skills, instructions, prompts) - Update migration trackers and status tables - Consolidate knowledge files following KISS/DRY principles ## When to Apply Trigger this skill when: - A session resolves a recurring problem (document the pattern) - A skill file is migrated or consolidated (update trackers) - A new workflow emerges (capture in appropriate .instructions.md) - User feedback reveals a principle ("don't over-simplify") ## Core Patterns ### Pattern Recognition Checklist Before session ends, ask: | Question | If Yes → Action | | -------- | --------------- | | Did we solve a problem that could recur? | Document in relevant skill file | | Did we learn something about the AI assistant's architecture? | Update relevant .github/ documentation | | Did a file get created/deleted/consolidated? | Update migration trackers | | Did user correct AI behavior? | Add to skill's Anti-Patterns + document principle | | Did the skill itself get improved during use? | Commit the refinement immediately | ### Documentation Location Guide | What You Learned | Where to Document | Audience | | ---------------- | ----------------- |