ccdawn-pr-review

Pass

Audited by Gen Agent Trust Hub on Jul 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a clear 'Context Boundary' and 'Allowed Action' set that restricts the agent to read-only operations. It explicitly forbids editing files, moving the HEAD, or performing merges and pushes.
  • [SAFE]: The instructions include safety-centric review criteria, such as checking for leaked credentials, privacy violations, and log leaks in the code under review.
  • [SAFE]: No evidence of prompt injection, obfuscation (Base64, zero-width characters, etc.), or suspicious external network communication was found.
  • [COMMAND_EXECUTION]: The skill allows for running read-only verification commands (e.g., tests, linting, builds). This behavior is consistent with the intended purpose of a PR review tool and is mitigated by the explicit prohibition of writing to the codebase or environment during the review process.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 13, 2026, 09:31 AM
Security Audit — agent-trust-hub — ccdawn-pr-review