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
Risk Level
SAFE
Analyzed
Feb 19, 2026, 04:11 AM