skill-smith

Creating specification-compliant Agent Skills involves fetching multi-page documentation and GitHub repositories to collect context, then organizing files with valid YAML frontmatter and progressive disclosure. This automated workflow ensures correct file organization and compliance with naming and size guidelines.

A valid Agent Skill package requires valid YAML frontmatter, a progressive disclosure structure with SKILL.md under 500 lines, and detailed content placed in references for on-demand loading. It must also comply with official Agent Skills naming and size guidelines before distribution.

Yes, you can use GitHub repositories to build AI agent skills by fetching and analyzing the repository contents alongside documentation sites and API references. This systematic resource gathering collects the necessary context and code samples to generate a spec-compliant skill package.

To validate an Agent Skill before sharing, you integrate with the official skills-ref library to check naming rules, frontmatter, and file structure. This built-in validation ensures the skill package meets v1.0 specification compliance for distribution.

Progressive disclosure in skill development is an organizational structure that keeps the main SKILL.md file under 500 lines by placing detailed content in a references directory for on-demand loading. This design ensures agents receive concise instructions while retaining access to extended context.

To refactor non-compliant agent skills to meet the v1.0 specification, you apply automated workflows that restructure files into progressive disclosure, validate YAML frontmatter, and move oversized content to references. This ensures the updated package passes naming and size guideline checks.