elysiajs

To extract backend metadata from YAML frontmatter, the Skill reads the SKILL.md file to parse the Skill's name and description. It then scans optional directories like scripts/ and assets/ to produce a machine-readable metadata payload for indexing workflows.

You can detect optional resource directories by scanning the repository structure for folders such as scripts/, references/, and assets/. This process identifies available resources and computes component dependencies to generate a complete metadata payload.

The best way to generate machine-readable metadata for backend indexing is to parse the YAML frontmatter from SKILL.md. This extracts core attributes and detects resource directories, producing a structured payload suitable for validation and documentation workflows.

Yes, you can compute dependencies, components, and toxicity status without external packages. The Skill analyzes the internal repository structure and YAML frontmatter directly to determine these metadata fields independently of external dependencies.

Skill Unit metadata extraction fails to find the description if the YAML frontmatter in SKILL.md is missing or malformed. Ensuring the frontmatter contains valid name and description fields is required for the parsing process to successfully generate the metadata payload.