review-pr
Pass
Audited by Gen Agent Trust Hub on May 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements a well-defined, read-only workflow for reviewing pull requests. It includes comprehensive 'Hard Rules' that explicitly prohibit destructive actions, such as file modifications, commits, or environment-mutating commands.
- [COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (
gh) to retrieve PR metadata and diffs. The usage of this tool is strictly bounded to non-mutative operations, which is enforced by the skill's internal instructions. - [DATA_EXFILTRATION]: While the skill accesses potentially sensitive PR data and code, there are no instructions or scripts that attempt to transmit this information to external or untrusted destinations. All analysis remains within the agent's processing context.
- [PROMPT_INJECTION]: The skill processes untrusted input in the form of PR titles, descriptions, and code diffs. This constitutes an indirect prompt injection surface. However, the threat is mitigated by the skill's strict read-only constraints and the absence of any tools that could be abused to achieve high-impact compromise.
Audit Metadata