CrowdStrike launches AI Kill Switch to stop and isolate rogue agents
Reported by Forbes, Stocktwits, CNBC and 2 more outlets
CrowdStrike added an AI Kill Switch to Falcon Guardian. The company says authorized administrators can stop suspicious agents, prevent them from relaunching and block their network access while investigating.
Key points
- Only authorized administrators can trigger the action, and each action is recorded in an audit trail
- CrowdStrike says its policy can prevent an agent from relaunching on a targeted host or host group
- The tool is being introduced in Falcon Guardian; CrowdStrike did not give pricing or broader availability details
Sources
- ForbesRead ↗CrowdStrike Looks To An AI Kill Switch To Beat Rogue Agents
- StocktwitsRead ↗CrowdStrike To Reportedly Buy AI Security Startups For $495M, Launches ‘Kill Switch’ For Rogue Agents
- CNBCRead ↗Jim Cramer likes CrowdStrike's new 'AI kill switch' and the cyber stock
- CNBCRead ↗CrowdStrike launches AI kill switch that it says can stop internal agents that go rogue
- CNBCRead ↗CrowdStrike's new kill switch can stop rogue AI agents in their tracks, CEO says
Details
TechRadar reports that Falcon Guardian already identifies approved and unauthorized AI agents, monitors their instructions and file access, and can detect prompt injection attacks. The Kill Switch adds a way for administrators to contain an agent they consider suspicious.
CrowdStrike says the Falcon sensor will first stop an agent from executing at the Gateway, with endpoint blocking planned soon. The company says the feature uses the sensor’s visibility into what an agent is told to do, communicates and executes.
Written from coverage as of
Images
- CNBC
- TechRadar
- Hacker News
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh