pr-comment
Pass
Audited by Gen Agent Trust Hub on Apr 20, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local commands using the
gh(GitHub) andjqCLI tools. This is the intended behavior for interacting with repository data and processing API responses. The script correctly avoids shell interpolation by passing arguments as arrays and usingjqto construct JSON payloads safely. - [DATA_EXFILTRATION]: Network activity is restricted to authenticated interactions with official GitHub API endpoints. No sensitive data, such as credentials or environment variables, is transmitted to unauthorized or third-party domains.
- [PROMPT_INJECTION]: The skill's instructions are focused on functional task orchestration and tone guidelines. There are no attempts to bypass safety filters, override system instructions, or extract sensitive model information.
- [SAFE]: The implementation includes several security-hardening features: a 'prepare' and 'submit' workflow that requires a user approval gate (
AskUserQuestion), a SHA-drift check to prevent posting to stale code, and a session-based integrity hash (payloadHash) that prevents tampering with the review data between the preview and submission steps.
Audit Metadata