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. Maintained by secureagentics on GitHub, where it has earned 523 stars from the community.
