dev-skill-authoring

Auditing SKILL.md metadata quality involves validating required fields (name, description, argument-hint) and checking progressive disclosure structure to ensure alignment with Anthropic-style best practices for discoverability and token efficiency.

Progressive disclosure in skill authoring structures SKILL.md content incrementally, optimizing token efficiency by ensuring scripts, references, and assets load only when needed according to the guide.

Creating a new agent skill involves describing desired functionality to generate a SKILL.md template with validated metadata fields, one-level-deep references, and structured progressive disclosure.

Yes, refining skill descriptions for discoverability involves tuning metadata fields and content structure to improve token efficiency while aligning SKILL.md with best-practice guidelines for metadata quality.

Validating references in SKILL.md files limits nested dependencies to exactly one level deep, ensuring scripts, references, and assets are used correctly according to structural guidelines.