skill-management

You audit agent skills by executing a 9-phase audit process that validates YAML frontmatter, verifies directory structure conventions, and detects manifest-implementation drift. This process classifies issue severity and provides validation scoring to ensure quality.

Progressive disclosure in agent skill development is a layered information approach that keeps SKILL.md files concise while storing detailed content in references. This enforces best practices by maintaining clean skill boundaries and preventing context overload.

You write discoverable descriptions for agent skills by following specific rules to include trigger phrases and explicit use cases. This ensures the description accurately represents processing logic and adheres to open standard metadata requirements.

Yes, you can use skill-management to scaffold new skills by generating the correct directory structure and mandatory SKILL.md files. This ensures your skill authoring starts with a compliant foundation before validation.

The YAML frontmatter requirements for agent skills mandate specific metadata fields that ensure discoverability and standard compliance. The skill-management skill validates these fields during the 9-phase audit to enforce open standard conventions.

Agent skill validation fails with manifest-implementation drift when declared metadata in the manifest does not match actual processing logic or components. The 9-phase audit process detects this drift and classifies it as a severity issue requiring correction.