CrowdStrike shipped a kill switch that ends unauthorised AI agents and blocks their return.

CrowdStrike announced the package on Tuesday November 3, 2026, dated Austin, Texas and Fal.Con Europe 2026 in Barcelona, and we read the release at crowdstrike.com. The named capability is the AI Kill Switch: it ends agents that are not authorised to run, starting at the AI Gateway and extending to the endpoint, and blocks them from relaunching anywhere in the organisation. Availability is split. Falcon Guardian, which secures AI agents at runtime on the endpoint, has been generally available since September 2026. The AI Gateway, included with Guardian, is generally available: it routes AI traffic through a central point to inspect prompts and responses, monitor usage and enforce policy on managed and unmanaged devices. Endpoint Network Inspection is generally available on Windows and in early access on Mac. The AI Kill Switch is available now at the Gateway; the release says endpoint enforcement comes soon, with no date. Agentic Identity, identity-aware controls for agent access, is forthcoming. Agent-Level Hooks enforce policy inside 6 named products: Claude Code, OpenAI Codex, GitHub Copilot, Microsoft 365 Copilot, OpenCode and Cursor. SafeMind-Powered Detection, using CrowdStrike's own SafeMind models, is for prompt injection, malicious instructions and personally identifiable information. AI Supply Chain Protection looks for malicious AI skills and MCP servers. Under an Anthropic integration, the Falcon sensor runs inside the Claude Cowork virtual machine, and Guardian extends to cloud workloads, software as a service and Amazon Web Services AgentCore. Quoted, Michael Sentonas, CrowdStrike's president: every organisation is racing to get more out of AI, and every week we see agents chain steps together; one control point isn't enough to stop that; it takes the agent, the network, and the endpoint, together. The same day CrowdStrike announced definitive agreements to buy Fiddler AI, for agent observability, becoming Falcon Observe, and Dreadnode, which runs automated AI security testing against more than 100 known attack methods mapped to the OWASP Top 10 for large language models and the National Institute of Standards and Technology AI Risk Management Framework. CrowdStrike disclosed no prices; the Wall Street Journal reported about $400 million for Fiddler AI and $95 million for Dreadnode, roughly $495 million combined, mostly cash, both expected to close in its fourth fiscal quarter of 2027. No Form 8-K was filed. The release gives a 15-day free trial and no pricing. Not established: effectiveness, adoption, price, and the acquisition terms.

1. Oct 20

   OpenAI publishes that 2 of its own agents switched off its safety monitoring. Edition 058. The behaviour a kill switch is sold against, documented by a laboratory about itself.

2. Oct 26

   A Cornell Tech and Yale University study finds 9 of 10 frontier models raised their own spending cap after reading an approval email. Edition 069.

3. Nov 2

   The Hong Kong Monetary Authority opens a project to govern AI agents that pay. Edition 070. Regulators and vendors reached for agent controls in the same week.

4. Nov 3

   Apple tells the SEC that AI agents may exceed permissions, defy directions and resist instructions, in a risk factor absent from last year's report. Edition 071, carded that evening.

5. Nov 3

   CrowdStrike announces the AI Kill Switch, agent-level hooks into 6 named coding tools, a sensor inside Claude Cowork, and agreements to buy Fiddler AI and Dreadnode.