Crowd Strike has launched an AI Kill Switch and other security controls for its Falcon Guardian platform to stop unauthorized activity of AI agents. The company’s new model combines agent-level hooks, endpoint network inspection and runtime controls to monitor activity of AI agents, and how they communicate and perform tasks.
Also Read: HENNGE One Launches SSO Integration with Yaritori
Its AI Gateway facilitates integrated prompt and response inspection, while its Safe Mind-powered detection looks for threat types including prompt injection, prompt hijacking and malicious instructions. It is also increasing protection of the AI supply chain, agent identities and the cloud environment. Its buyout of Fiddler AI, for additional agent observability, is designed to give enterprises the ability to assess and monitor AI systems before deployment.


