Agent runtime security fails for agents that bypass enterprise gateways

Vendors often claim agent runtime security can block bad agent actions in real time. The article argues this holds only where agent traffic passes through an enterprise-owned gateway, which most agents never do. It cites a recent case of an employee using an AI assistant to trigger a mass-email automation, but the excerpt is truncated and gives no outcome.

For IT people · Written by AI from 1 source · Wrong?