code-review
Pass
Audited by Gen Agent Trust Hub on Nov 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
coderabbitCLI andgitcommands to perform its core functions. It implements defensive execution patterns, requiring the agent to verify the tool's absolute path and authentication status (viacoderabbit auth status --agent) before running any review commands. It also provides instructions for handling sandbox environments using theagyCLI in a user-mediated recovery flow, ensuring the agent does not attempt to bypass permissions autonomously. - [INDIRECT_PROMPT_INJECTION]: The skill identifies a potential attack surface where malicious content in code repositories or external API responses could influence the agent's behavior. It mitigates this by instructing the agent to treat all repository data and CLI output as untrusted and requiring explicit user confirmation before executing any instructions or code found in those sources.
- [EXTERNAL_DOWNLOADS]: The skill uses the CodeRabbit CLI to send code diffs to the CodeRabbit API for analysis. This is documented as a standard part of the service, and the instructions advise against including files with sensitive credentials in the review scope to prevent accidental data exposure.
Audit Metadata