book-translation
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- Prompt Injection (SAFE): No instructions were found that attempt to override agent behavior or bypass safety guidelines.\n- Data Exposure & Exfiltration (SAFE): Access is limited to project files required for translation. No sensitive system paths, credentials, or unauthorized external network calls were detected.\n- Unverifiable Dependencies & Remote Code Execution (SAFE): The skill suggests running local scripts (
node,npm) already present in the repository, which is standard for development workflows. No external script downloads or remote code executions were found.\n- Indirect Prompt Injection (SAFE): The skill processes project content for translation but does not provide a surface for executing instructions embedded in that data or granting excessive capabilities.
Audit Metadata