ai-wrapper-product
Pass
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No instructions found that attempt to override agent behavior or bypass safety guidelines.\n- [Data Exposure & Exfiltration] (SAFE): No sensitive data exposure or exfiltration patterns detected. The code uses placeholders for context and user data.\n- [Unverifiable Dependencies & Remote Code Execution] (SAFE): References standard libraries like @anthropic-ai/sdk. No remote script execution or piped downloads found.\n- [Indirect Prompt Injection] (SAFE): The skill provides patterns for handling user input that include validation and the use of structured SDK message roles, which are defensive measures against injection.
Audit Metadata