The Agent Skills Directory

[SAFE]: The skill is a documentation and interface definition for a code validation tool. It does not contain executable malicious code, hardcoded credentials, or unauthorized network operations. Its purpose is to improve the security and reliability of other scripts.
[PROMPT_INJECTION]: The skill processes content from external markdown files (e.g., under plugins/*/commands/*.md), which represents an indirect prompt injection surface.
Ingestion points: Markdown files passed as arguments or located in plugin directories.
Boundary markers: Not specified in the instructions.
Capability inventory: Shell syntax validation (bash -n, shellcheck) and potential runtime execution of 'GREEN-tier' commands.
Sanitization: Not explicitly mentioned, though the tool itself is designed to flag unsafe patterns.
[COMMAND_EXECUTION]: The skill description indicates that the validation tool may execute 'GREEN-tier' read-only commands at runtime to detect failures like broken jq filters or invalid grep patterns. This is a functional requirement of the validator and is performed within a development context.

validate-skills