Open-source runtime AI agent security tool - monitors and controls AI agents, catching malicious tool use, prompt injection, and policy drift in real time, before the agent acts.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).