Prompt Inspector

Detect prompt injection attacks and adversarial inputs in user text before passing to LLM