Keeping AI Agents on Task: Check Point Advances AI Agent Safety With NVIDIA OpenShell Reference Architecture
Check Point, Wednesday, October 21st, 2026
Check Point describes its NVIDIA OpenShell work on agent governance, runtime control and in-silicon security for AI agents.
Check Point says AI agents now plan, use tools, work with sensitive data and hand work to other agents, so an agent can follow every rule and still do the wrong thing.
Isolation, network security and monitoring still matter, but an allowed action can still be wrong.
Check Point describes its collaboration with NVIDIA on the OpenShell Reference Architecture.
The work covers agent governance, control over how agents run, and in-silicon security built into AI infrastructure.