Smart Restaurant OS — Skill Layer (Domain Intelligence for AI Agents)

To validate YAML frontmatter in SKILL.md files, scan the repository root to identify Skill Units and verify the presence of required metadata fields like name and description. The extractor checks each file and flags invalid or unsafe content before activation.

The best way to extract metadata from Skill Units is to scan the repository root and run an automated metadata extractor. This process reads YAML frontmatter to pull dependencies, components, category, and keywords, producing a structured YAML payload for centralized search.

You can flag unsafe or missing frontmatter by verifying the SKILL.md file at the root of each Skill Unit. The validation process isolates invalid or unsafe Skill Units that lack proper frontmatter or contain risky content, preventing them from entering the discovery catalog.

Yes, you can generate a structured YAML payload for downstream tooling by scanning valid Skill Units. The extractor compiles extracted metadata such as name, description, dependencies, and category into a structured format ready for centralized search and discovery.

AI governance metadata extraction works for repositories without dependencies because the validator operates independently on SKILL.md files. It checks frontmatter presence and extracts available metadata fields, requiring no external dependencies to produce the structured YAML payload.