Safety
Add Runtime Controls to AI Agents with NVIDIA OpenShell
NVIDIA OpenShell 0.1.0 is an open‑source runtime that enforces which systems and data an AI agent may access without requiring changes to the agent itself. Its components—Gateway, Supervisor, Sandbox,
NVIDIA OpenShell 0.1.0 is an open‑source runtime that enforces which systems and data an AI agent may access without requiring changes to the agent itself. Its components—Gateway, Supervisor, Sandbox, and a formal policy engine—provide sandboxed execution, controlled service API permissions, credential isolation, and model‑checked policy enforcement across Docker or Kubernetes environments. Companies such as Cadence, Slack and Gecko Robotics deploy OpenShell to secure chip design, enterprise automation and physical robotics workloads while enabling shared compute workspaces.
Related
- NVIDIA Open Agent Safety Platform: A Reference for Continuous In-Silicon Agent Monitoring
- Control agent behaviors and cost beyond a single action: new capabilities in Amazon Bedrock AgentCore
- Secure AI agents with Policy and Lambda interceptors in Amazon Bedrock AgentCore gateway
- How DigitalOcean Manages Credentials for Autonomous Agents
Source: NVIDIA Developer | 2026-09-28