visual-review

Pass

Audited by Gen Agent Trust Hub on Aug 23, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses gh pr view, gh pr diff, and gh auth status to fetch pull request information and diffs. These are standard operations for a developer tool designed for GitHub integration. It also uses open (macOS) or xdg-open (Linux) to display the generated local HTML report, which is expected behavior for its primary purpose.
  • [DATA_EXPOSURE_&_EXFILTRATION]: No sensitive data access or network exfiltration to untrusted domains was detected. All network communication is performed via the GitHub CLI (gh) to official GitHub endpoints. The resulting HTML report is generated locally and intended for offline use, explicitly stating a 'no CDN' and 'self-contained' policy to avoid external leaks.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted pull request diffs and file content. It has an ingestion point in Phase 2 where it reads changed files. However, the capability surface is limited to generating a local HTML file and opening it in a browser. There are no automated write-back actions to the repository or remote execution of the analyzed code. The principles in references/principles.md act as a narrow reasoning lens to further constrain the AI's assessment of the data.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 23, 2026, 06:03 PM
Security Audit — agent-trust-hub — visual-review