Uploaded September 2026 | Updated November 2026, 3 hours ago
Learn how to place an AI agent you already run inside NVIDIA OpenShell without rewriting the agent or changing its harness. This tutorial shows how NVIDIA OpenShell creates a governed runtime boundary around the agent, its tools, Model Context Protocol (MCP) servers, and delegated subagents, so every component inherits consistent policy, credential, and audit controls from launch.
The walkthrough begins with an existing agent and shows the correct control sequence: an orchestrator creates an OpenShell runtime, then the harness starts inside it. You’ll see how to configure filesystem, network, and process policies at the level of binaries, destinations, methods, and paths. The demo then deliberately triggers a blocked action, follows the agent’s request for a scoped policy change, and shows how an approved update takes effect without restarting the sandbox.
Next, the tutorial covers policy advising with auto-approvals. You’ll see how upper-limit ceilings let routine requests resolve automatically while keeping the agent from granting itself broader authority. The same model extends to long-running agents that create short-lived subagents for individual tasks: child runtimes inherit controls and cannot exceed the parent’s limits.
The final section shows how NVIDIA OpenShell supports shared production environments. It introduces sandbox grouping for teams and business units, Kubernetes deployment and high-availability choices, centralized sandbox logs, and the audit records operators need to understand what ran, what was allowed, what was denied, and what changed. It also explains where extensibility APIs fit when organizations need custom policy parsers, compute drivers, credential stores, or third-party service integrations.
By the end, you’ll have a practical pattern for adding sandboxed execution, live policy updates, delegated-agent controls, and observability to an existing agent stack.
➡️ Learn more:
📝 Tech blog: developer.nvidia.com/blog/add-runtime-controls-to-ai-agents-with-nvidia-openshell
📥 GitHub repository:github.com/NVIDIA/OpenShell
📘 Build: build.nvidia.com/openshell
Learn how to place an AI agent you already run inside NVIDIA OpenShell without rewriting the agent or changing its harness. This tutorial shows how NVIDIA OpenShell creates a governed runtime boundary around the agent, its tools, Model Context Protocol (MCP) servers, and delegated subagents, so every component inherits consistent policy, credential, and audit controls from launch.
The walkthrough begins with an existing agent and shows the correct control sequence: an orchestrator creates an OpenShell runtime, then the harness starts inside it. You’ll see how to configure filesystem, network, and process policies at the level of binaries, destinations, methods, and paths. The demo then deliberately triggers a blocked action, follows the agent’s request for a scoped policy change, and shows how an approved update takes effect without restarting the sandbox.
Next, the tutorial covers policy advising with auto-approvals. You’ll see how upper-limit ceilings let routine requests resolve automatically while keeping the agent from granting itself broader authority. The same model extends to long-running agents that create short-lived subagents for individual tasks: child runtimes inherit controls and cannot exceed the parent’s limits.
The final section shows how NVIDIA OpenShell supports shared production environments. It introduces sandbox grouping for teams and business units, Kubernetes deployment and high-availability choices, centralized sandbox logs, and the audit records operators need to understand what ran, what was allowed, what was denied, and what changed. It also explains where extensibility APIs fit when organizations need custom policy parsers, compute drivers, credential stores, or third-party service integrations.
By the end, you’ll have a practical pattern for adding sandboxed execution, live policy updates, delegated-agent controls, and observability to an existing agent stack.
➡️ Learn more:
📝 Tech blog: developer.nvidia.com/blog/add-runtime-controls-to-ai-agents-with-nvidia-openshell
📥 GitHub repository:github.com/NVIDIA/OpenShell
📘 Build: build.nvidia.com/openshell










