The Agent Skills Directory

[SAFE]: The skill's functionality is limited to static analysis of local source code. It uses built-in tools (Read, Grep, Glob) to facilitate the audit process without requiring elevated privileges or network access.
[PROMPT_INJECTION]: The skill evaluates content from external files provided via $ARGUMENTS, which represents an indirect prompt injection surface.
Ingestion points: Code and documentation files at the path specified in $ARGUMENTS (SKILL.md).
Boundary markers: Absent; the skill does not explicitly instruct the agent to ignore instructions embedded within the code being audited.
Capability inventory: File system read access via Read, Grep, and Glob tools. No shell execution or network capabilities are permitted in the configuration.
Sanitization: None; the skill is designed to interpret the content of the files as source code for compliance checking.

a11y-audit