prompt-guard

Pass

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's primary purpose is to document the use of Meta's Prompt Guard security classifier. The instructions and code provide legitimate utility for developers building LLM applications.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the HuggingFace model hub to download the 'meta-llama/Prompt-Guard-86M' model. HuggingFace is a well-known service, and the model is published by a trusted organization.
  • [PROMPT_INJECTION]: The markdown documentation contains several examples of common prompt injection and jailbreak phrases (e.g., 'Ignore all previous instructions', 'Pretend you're a pirate'). These are used exclusively as test cases for the classifier and are not instructions directed at the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 29, 2026, 12:55 AM
Security Audit — agent-trust-hub — prompt-guard