The Agent Skills Directory

[SAFE]: The skill functions as a development tool to audit and validate other skills. It uses a local script to perform structural and semantic checks.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute its own local linting script (scripts/skill-lint.js). This execution is restricted to node commands as defined in the allowed-tools frontmatter, which is a secure practice for this context.
[DATA_EXPOSURE]: The linting script reads project files (skills, agents, and references) to verify their contents. The gathered data is used solely to generate a health report for the user and is not exfiltrated to any external domains.
[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from other skill files. The processing is restricted to structural analysis (e.g., regex checks for headers, line counting, and Jaccard similarity for descriptions) and does not involve executing instructions contained within the analyzed data. The risk surface is minimized by the analytical nature of the script.

skill-health-check