ai-patterns-tool-use-patterns

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a documentation and pattern library for developers. It does not execute any dangerous commands or scripts.
  • [SAFE]: The skill explicitly promotes security best practices, including the mandatory use of Zod or JSON Schema to validate LLM-generated arguments, which are treated as untrusted input.
  • [SAFE]: The instructions include 'Critical Requirements' for guarding loops with maximum step limits to prevent runaway costs or resource exhaustion.
  • [SAFE]: The 'Advanced Examples' provide defensive patterns such as domain allowlisting for URL-based tools and human-in-the-loop approval gates for destructive actions (e.g., database deletions or fund transfers).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 05:41 PM
Security Audit — agent-trust-hub — ai-patterns-tool-use-patterns