guardrails
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is purely instructional and defensive in nature. It defines a layered circuit breaker model (Intent, Input, Tool, Output, and Audit) to improve agent safety.
- [SAFE]: The content specifically addresses and provides mitigations for security risks such as prompt injection, sensitive data disclosure, and insecure tool design.
- [SAFE]: All external references in the metadata point to trusted and well-known documentation sources, including OpenAI, NIST, and OWASP.
- [SAFE]: The skill correctly identifies protected file paths (e.g., .env, .pem, id_rsa) and secret patterns (e.g., sk_live_, ghp_) to instruct the agent on what to protect, rather than attempting to access them maliciously.
Audit Metadata