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
Risk Level
SAFE
Analyzed
Feb 17, 2026, 04:49 PM