llamaguard
Pass
Audited by Gen Agent Trust Hub on Jun 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill facilitates the use of Meta's LlamaGuard models hosted on HuggingFace. As these resources originate from a well-known and trusted organization, the external references are considered safe.
- [SAFE]: The required dependencies (transformers, torch, vllm) are standard, reputable open-source libraries within the machine learning ecosystem.
- [SAFE]: The provided code implementations utilize
apply_chat_templateto format conversation data. This approach establishes clear structural boundaries between user input and model instructions, which is a recommended practice to mitigate risks associated with indirect prompt injection.
Audit Metadata