llamaguard
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a technical guide for implementing AI safety guardrails. It references official and well-known model repositories from Meta's LlamaGuard project on HuggingFace. All documented dependencies are standard industry-standard libraries (transformers, torch, vllm, fastapi). The workflows describe standard content moderation patterns without any evidence of malicious instructions, data exfiltration, or obfuscation.
Audit Metadata