review-implementation
Pass
Audited by Gen Agent Trust Hub on Jul 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill ingests untrusted implementation diffs and specification files, which provides a surface for indirect prompt injection.
- Ingestion points: Implementation references (git refs, diff text) and specification files (spec.md) are read from the repository or user input.
- Boundary markers: The instructions do not specify any delimiters or safety markers to differentiate untrusted data from the agent's internal logic.
- Capability inventory: The skill reads local files, executes git diff, and writes review reports to the local filesystem; it is explicitly prohibited from executing the code it reviews.
- Sanitization: No sanitization or content validation mechanisms are defined for the inputs.
Audit Metadata