self-improvement

Continuous improvement is achieved by capturing command failures, exceptions, and insights in a structured format. The skill automatically detects errors via hooks and supports manual logging for developers to systematically record interactions.

Promoting learnings to permanent project memory involves moving structured logs into files like CLAUDE.md or AGENTS.md. This facilitates capturing corrections and knowledge gaps permanently for future project reference.

Skill extraction from recurring errors uses a structured workflow to turn valuable learnings into reusable skills. By reviewing logged command failures and user corrections, developers can create new skills to prevent similar problems.

Automatic error logging via hooks records command failures, exceptions, and unexpected behaviors without manual intervention. This mechanism systematically captures errors encountered during development interactions for later review.

Tracking feature requests and knowledge gaps requires documenting user-requested capabilities and logging corrections from interactions. This structured approach ensures insights are captured for future bug fixes and documentation updates.

Automatic logging for continuous improvement may not suit contexts requiring manual curation of learnings or when hook-based error detection conflicts with existing project memory workflows. Manual logging provides more control over captured insights.