skill-refine

To audit a Claude skill against Anthropic best-practices standards, the skill validates frontmatter, structural hygiene, references, and domain-depth against the official checklist to ensure compliance and reliability.

The best way to refine a Claude skill while preserving its intent is to apply a phase-by-phase workflow. This proposes concrete, minimally invasive changes to improve structure and domain-depth without altering the original purpose.

Check Claude skill frontmatter and references by running a Phase 1-5 refinement process on the SKILL.md. This validates structural hygiene and produces concrete improvement recommendations for checklist compliance.

Yes, you can add domain-research to an existing Claude skill to improve reliability. The refinement process proposes domain-research additions and concrete improvements to keep skills current with Anthropic standards.

Claude skill refinement requires no specific dependencies. It operates directly on the target skill's SKILL.md and references, auditing them against the official best-practices checklist to produce actionable improvements.

A Claude skill fails the Anthropic best-practices audit due to frontmatter errors, poor structural hygiene, missing references, or insufficient domain-depth. The refinement process identifies these gaps and proposes minimally invasive fixes.