koa-typescript

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION] (SAFE): The skill uses persona-setting language ('You are an expert...') to define the agent's role but does not contain any patterns aimed at bypassing safety filters or overriding system instructions.
  • [CREDENTIALS_UNSAFE] (SAFE): No hardcoded credentials were found. The skill explicitly recommends using environment variables for secrets (JWT_SECRET) and warns against committing secrets to version control.
  • [EXTERNAL_DOWNLOADS] (SAFE): The documentation and code snippets reference standard, well-maintained libraries from the Node.js ecosystem (e.g., Koa, Zod, Helmet).
  • [DATA_EXFILTRATION] (SAFE): No unauthorized data access or network transmission patterns were detected. Network-related code is limited to standard middleware configuration.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:29 PM