🛡️ PromptGuard

LLM prompt safety evaluator — powered by RAG + Llama
Built on a 181k prompt safety dataset.

Try an example