Red Hat launches OpenShell for secure AI agent control

Red Hat just announced OpenShell, an open-source runtime that keeps AI agents in check. It's built to help big companies deploy agents that can write code or access internal systems—without risking security disasters. OpenShell enforces strict rules on what agents can do, tracks every action, and lets humans approve any big changes. NVIDIA's involved too, baking OpenShell into their new Open Agent Safety Platform.

• OpenShell is open source and enforces security by default
• Checks agents' actions before allowing system changes
• Supports human approval for risky agent behaviors
• Integrated with NVIDIA's Open Agent Safety Platform
• Built for enterprise-scale deployments, not DIY fixes