The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were identified. The skill is a local documentation auditor that evaluates skill quality across several dimensions. It does not perform network requests, execute external scripts, or access sensitive system files.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it is designed to ingest and process the full content of external SKILL.md files for quality auditing. This is a characteristic of its primary function.
Ingestion points: The skill reads the content of SKILL.md and reference files from directories specified in the repository or user input.
Boundary markers: The ingested text is processed without explicit delimiters or instructions to the agent to disregard embedded commands during analysis.
Capability inventory: The skill's capabilities are limited to local file system enumeration and read operations; no write or network capabilities are present.
Sanitization: The skill does not implement sanitization or filtering of the Markdown content before evaluation.

skill-evaluator