The Agent Skills Directory

[SAFE]: The skill executes local repository scripts (plugin-compliance-check.sh, audit-skill-descriptions.py, audit-skill-structure.py) to perform its audit. This behavior is consistent with its stated purpose as a health-check tool and does not involve external downloads or remote code execution.
[SAFE]: Employs the dynamic context injection syntax (exclamation mark followed by backticks) to run find commands at load time for project introspection. These commands are used to locate scripts and previous audit results and do not target sensitive system files or credentials.
[SAFE]: The skill processes other skill files and generates reports, creating an indirect prompt injection surface. This is considered acceptable as it operates on local repository content and is core to its auditing function.
Ingestion points: Analyzes local SKILL.md files and reads generated reports in the tmp/skill-audit/ directory.
Boundary markers: No explicit delimiters or instructions are used to separate the analyzed content from the agent's instructions.
Capability inventory: Uses the Bash, Read, and TodoWrite tools as defined in the skill frontmatter.
Sanitization: The process involves human editorial review of findings rather than automated content sanitization.

health-skill-audit