github-code-review
Warn
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: MEDIUMCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill attempts to extract GitHub authentication tokens by reading the
~/.git-credentialsfile. - [EXTERNAL_DOWNLOADS]: Communicates with the official GitHub REST API (
api.github.com) to retrieve pull request metadata and post review comments. - [COMMAND_EXECUTION]: Uses
python3for processing JSON data and executes various CLI tools includinggit,gh,pytest, andrufffor repository analysis and testing. - [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests untrusted code changes from git diffs and pull requests.
- Ingestion points: External pull request content and file diffs analyzed during the review process.
- Boundary markers: No explicit delimiters or warnings are implemented to prevent the agent from following instructions embedded in the code being reviewed.
- Capability inventory: The skill possesses capabilities for network operations (
curl), system commands (git,gh), and script execution (python3). - Sanitization: No sanitization or validation of the ingested code content is performed before analysis.
Audit Metadata