skill-improver

Auditing LLM skills involves validating frontmatter, activation contracts, decision gates, and execution workflows to identify structural problems and compliance gaps in SKILL.md files. This process generates actionable quality reports for agent skill libraries.

Refactoring AI agent skills safely requires preserving author intent, activation behavior, and important constraints while normalizing SKILL.md files. Structured quality audits identify instruction clarity issues and suggest improvements before applying changes.

Validating skill frontmatter and execution workflows requires checking metadata compliance, decision logic, and output contracts against established standards. This detects structural issues and unclear instructions in existing LLM skill definitions.

Automated quality reviews can normalize SKILL.md files by auditing metadata, triggers, and rules to produce actionable quality reports. This identifies compliance gaps and suggests improvements while maintaining the original skill constraints.

AI agent workflows fail when skill definitions contain structural issues, unclear instructions, or metadata compliance gaps. Auditing SKILL.md files validates output contracts and decision gates to resolve these execution problems.

Before refactoring LLM skill definitions, check skill metadata, triggers, structure, rules, and output contracts against established standards. This safe refactoring guidance preserves author intent while identifying compliance gaps to fix.