As AI agents move from generating answers to executing code and accessing business systems, the challenge increasingly moves beyond just running day-to-day operations. 

A primary customer focus today is controlling what autonomous agents are allowed to do. That brings us to OpenShell. It's an open-source secure agent runtime that Red Hat is helping develop alongside NVIDIA and, of course, the broader community. 

The approach puts enforcement into the environment where the agent operates, of course, setting boundaries around what it can access and execute rather than relying on a model alone to follow the rules. 

We think this is a central question for the next chapter of Cloud Native.