big-review
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE]: The skill retrieves code diffs and full file contents from local git repositories and remote GitHub repositories. This behavior is the primary intended function of the skill and uses official command-line tools ('git' and 'gh') for read-only context gathering.
- [COMMAND_EXECUTION]: The skill provides instructions for executing shell commands such as 'git merge-base', 'git diff', and 'gh api' to fetch repository metadata and contents. These commands are necessary for performing code reviews and are explicitly directed to be used in a read-only manner unless the user requests a fix or a review post.
- [INDIRECT_PROMPT_INJECTION]: The skill's core function involves processing untrusted data in the form of code from pull requests or local branches.
- Ingestion points: Code content is ingested from local diffs ('git diff') and GitHub pull requests ('gh pr diff', 'gh api').
- Boundary markers: The skill relies on an 'intent pass' and 'verification gate' to validate findings but does not define explicit technical delimiters for isolating untrusted code instructions from the agent's logic.
- Capability inventory: The skill can execute shell commands and interact with the GitHub API to post comments or requests for changes.
- Sanitization: The skill focuses on semantic and correctness analysis; no specific sanitization of the code being reviewed is mentioned before it is processed by the agent.
Audit Metadata