The agentic kill switch
An autonomous agent doesn't get to finish what you didn't approve. Flip it and the agent's next action is refused mid-run, so the step it was about to take is the step that never happens.
Two switches, actually. One stops an autonomous agent before its next action. The other halts AI across every enrolled device and provider in one move. Both immediate, both reversible, both recorded.



Sometimes the problem is one agent: it's autonomous, it's mid-run, and it's about to take a step you never approved. Sometimes the problem is wider: a compromised tool, an injected instruction, or a decision to pause AI while you investigate. You shouldn't have to chase either one down device by device, so Verillian gives you a stop for each. Neither waits on a vendor.
A stop you'd hesitate to use is no stop at all. These two are built to be pulled without ceremony and cleared without cleanup.
An autonomous agent doesn't get to finish what you didn't approve. Flip it and the agent's next action is refused mid-run, so the step it was about to take is the step that never happens.
One move halts every connection through the checkpoint, on every enrolled device and for every provider, with no parsing in the way and no vendor's cooperation needed. The tools nobody registered go quiet with everything else.
The activation itself lands in the signed record, and so does the moment you bring things back. Later, you can show exactly when AI stopped, who stopped it, and when it resumed.
Halting a fleet shouldn't be a stray click, so the console makes you type the confirmation before anything stops. Clearing it works the same deliberate way, and both moments are sealed the instant they happen.

Watch the whole arc: traffic flowing, containment landing, the fleet coming back. Each next connection is refused at the checkpoint the moment containment lands, and every transition seals its own entry in the chain.
The goal is not to stop AI. It is the opposite, to let institutions finally say yes.