Skill Evolution Manager

You can automatically extract skill improvements from conversation logs by analyzing completed dialogues to surface actionable feedback and persisting it for future iterations. This process derives structured updates from user outcomes and session reviews.

The best way to update SKILL.md with learned best practices while preserving history is to use smart document stitching. This technique integrates new insights directly into the skill file without overwriting previous versions.

Storing evolution data in evolution.json helps with skill version control by providing a structured record of extracted feedback. This allows alignment scripts to apply accumulated insights to new skill releases consistently.

Yes, you can perform cross-skill review and performance diagnostics after a session. The workflow analyzes completed conversations across multiple skills to identify actionable improvements and guide their iterative evolution.

No external dependencies are required to run the skill evolution workflow. The process relies entirely on internal scripts to analyze dialogues, generate evolution.json, and stitch updates into SKILL.md.

You should trigger the skill evolution workflow to align versions across releases after completing a conversation session. This ensures that newly derived insights are captured, persisted, and applied to new skill releases via alignment scripts.