tool-use-guardian
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: Procedural Content. The skill consists entirely of high-level instructions for the agent to follow when a tool fails or times out. It does not include scripts, shell commands, or automated routines.
- [SAFE]: Data Privacy. The instructions explicitly direct the agent to 'not log credentials, private payloads or full conversation context' and to not change credentials to evade failures, which aligns with security best practices.
- [SAFE]: External References. The declared source links point to the author's GitHub repository and informational pages on Vercel. These are listed for reference and are not involved in any automated downloads or code execution.
- [SAFE]: Input Handling. While the skill processes external tool outputs (a potential surface for indirect prompt injection), the instructions mitigate this by requiring validation against actual tool schemas and prohibiting the 'guessing' or 'coercion' of malformed data.
Audit Metadata