What problem does it solve?
When source code changes, existing AI skills become stale and inaccurate. Manually regenerating skills discards valuable developer customizations and annotations. This Skill automates the surgical update of skills, ensuring only changed exports are refreshed while all developer-authored [MANUAL] sections remain intact.
Core Features & Use Cases
- Surgical Regeneration: Only changed exports are re-extracted using AST-backed analysis; unchanged content is never touched, preserving context and reducing noise.
- [MANUAL] Section Preservation: Developer-authored content blocks within generated SKILL.md files survive regeneration with zero content loss, maintaining custom examples and notes.
- Provenance Tracking: Every regenerated instruction traces to its source code with file:line citations, ensuring full traceability and auditability.
- Use Case: A developer modifies a public API by adding a new parameter. This Skill updates only the affected function's documentation in the skill file while keeping all custom usage examples and warnings intact.
Quick Start
Use the skf-update-skill workflow to surgically update an existing skill after source code changes while preserving all [MANUAL] sections.