The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted codebase data to generate health scores and reports, which represents a potential injection surface.
Ingestion points: health-guide.md instructs the agent to read top-level structure, entry points, import statements, diffs, and test files.
Boundary markers: None explicitly mentioned in the provided guide files.
Capability inventory: Analysis shows only read operations and report generation; there are no network operations, file-system writes, or subprocess execution capabilities defined.
Sanitization: Not explicitly detailed in the orchestration process.
[SAFE]: No malicious patterns were identified. The skill follows established patterns for codebase linting and auditing. All identified behaviors, such as reading project metadata and source code, are strictly limited to the skill's stated purpose of quality assessment.

brooks-health