respond-to-code-review
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses platform-specific CLI tools, such as
ghfor GitHub andglabfor GitLab, to fetch pull request data, manage review states, and post consolidated responses. - [EXTERNAL_DOWNLOADS]: The skill includes procedures to download and install the GitHub CLI from its official GitHub repository if it is not already available in the environment.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from pull request comments and discussion threads, which could potentially contain malicious instructions intended to manipulate the agent's behavior when it applies code fixes or generates replies.
- Ingestion points: Feedback is collected through API calls and CLI commands targeting pull request comments, inline discussions, and activity logs on GitHub, GitLab, and Bitbucket.
- Boundary markers: The skill instructs the agent to 'Verify claims before changing code or replying', which serves as an instructional verification step, although no specific technical delimiters are implemented to isolate external content.
- Capability inventory: The skill is capable of performing file system writes to address review feedback, executing shell commands for git and platform operations, and initiating network requests via established CLI tools.
- Sanitization: The instructions do not define specific sanitization or filtering logic for the content of the review comments before they are processed by the agent.
Audit Metadata