Jev as an agent safety monitor

The claim

Checking each agent action with Jev first catches most attacks with almost no false blocks, at much lower latency.

Our read

“Most” and “almost no” are doing all the work. No attack set, no counts, no false-positive rate. The pattern is sound; the result isn't checkable.

Unverified: A specific claim with nothing public to check it against.

Test status

We haven't reproduced this yet. When we do, measured accuracy, latency, and cost will appear here next to the claimed figures.