autoreview
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates the execution of external review engines and Git commands. Technical analysis of the provided scripts shows that these operations are conducted within isolated temporary workspaces with sanitized environments to protect the host system.
- [DATA_EXFILTRATION]: While the skill transmits code to external AI providers, it incorporates extensive validation logic to detect and block secrets (such as API keys and private keys) and sensitive file paths (like .env or .ssh/ credentials) before they are processed by the review engines.
- [REMOTE_CODE_EXECUTION]: The skill invokes remote model analysis through restricted official CLIs. These interfaces are configured to disable potentially dangerous local tools and host filesystem access, ensuring the review process remains confined to the provided code bundle.
- [COMMAND_EXECUTION]: The diagnostic and testing scripts included with the skill use subprocess execution to verify functionality. These executions are limited to controlled test fixtures and temporary directories, following standard development practices for security-critical tooling.
Audit Metadata