receiving-code-review
Pass
Audited by Gen Agent Trust Hub on Apr 8, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No patterns indicative of prompt injection or instructions to bypass safety guidelines were detected in the skill content.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill does not access sensitive files or hardcode credentials, nor does it perform network operations.
- [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: There are no instructions for downloading remote scripts or installing external dependencies.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external feedback from code reviews (ingestion point). Although specific delimiters or sanitization are not defined, the skill mitigates risk by instructing the agent to independently verify and test suggestions using its available file and shell tools (capability inventory) before making changes.
Audit Metadata