How to use Open Shell (NVIDIA/OpenShell): 3 use cases

3 use cases

Run a coding agent in a locked-down sandbox

Install Open Shell, create a sandbox and run an agent that must ask before it reaches anything new.

View prompt

Install NVIDIA/OpenShell on this machine, create a sandbox named <name> with `openshell sandbox create`, and run <agent> inside it on <task>. Every time it asks for new file or network access, show me the request and wait for my approval before allowing it.

Let agents use API keys they can never see

Give agents credentials that only work at approved endpoints, so a leaked prompt cannot leak the key.

View prompt

Using NVIDIA/OpenShell, set up a provider for <service> so the agent in sandbox <name> can call <approved endpoint> without ever reading the real key. Then try a request to an unapproved host and confirm it is blocked, and show me what the policy prover flags if I add that host.

Teach your coding agent to drive Open Shell

Install the official Open Shell skills so your agent can run the CLI, write policies and debug gateways.
- REPO NVIDIA/OpenShell
- Claude Code
- Codex
- Cursor
- Software and DevOps
- Intermediate
View prompt
```
Install the NVIDIA/OpenShell skills with `npx skills add NVIDIA/OpenShell`, then write a sandbox policy for <project> that allows <files and hosts> and nothing else. Explain each rule, and do not apply it until I approve.
```

Quick start with Open Shell

Check you have what it needs: Linux, macOS on Apple Silicon, or Windows with WSL 2 (experimental); Docker, Podman, or host virtualization; An agent to run inside the sandbox; the default image is minimal Ubuntu with no agent installed.
Open NVIDIA/OpenShell on GitHub and read the README and license.
Run curl -LsSf https://raw.githubusercontent.com/NVIDIA/OpenShell/main/install.sh | sh to install the CLI and a local gateway
Run openshell sandbox create --name demo to create your first sandbox
Follow the Run Your First Agent guide in the docs, which runs OpenCode against a free OpenRouter model inside the sandbox
When the agent needs new access, approve it from Open Shell instead of granting it up front

These steps come from our top use case, Run a coding agent in a locked-down sandbox, which also has a copy-paste agent prompt.

Use this in your AI agent

Ask your agent

Using the Deduped MCP, get the setup steps for NVIDIA/OpenShell

Add Deduped to Claude Code once

claude mcp add --transport http deduped https://dedupedai.com/mcp

Cursor, ChatGPT, others

Open Shell alternatives · Best AI repos for Software and DevOps

About this repo

Who it is forDevelopers and platform teams who want coding agents to install packages, call APIs and use credentials without giving them the whole machine.
Why it is hot20k stars, picked up by one AI source we track
LanguageRust
Sourcegithub.com/NVIDIA/OpenShell · Back to Hot repos

Requirements

Linux, macOS on Apple Silicon, or Windows with WSL 2 (experimental)
Docker, Podman, or host virtualization
An agent to run inside the sandbox; the default image is minimal Ubuntu with no agent installed
For Kubernetes: Helm and a CNI that enforces NetworkPolicy

Add this badge to your README

Maintain Open Shell? Show readers where to find these use cases.

Markdown

[![Use cases on Deduped](https://dedupedai.com/badge/NVIDIA/OpenShell.svg)](https://dedupedai.com/use-cases/nvidia-openshell/?utm_source=github&utm_medium=badge&utm_campaign=NVIDIA-OpenShell)

HTML

<a href="https://dedupedai.com/use-cases/nvidia-openshell/?utm_source=github&amp;utm_medium=badge&amp;utm_campaign=NVIDIA-OpenShell"><img src="https://dedupedai.com/badge/NVIDIA/OpenShell.svg" alt="Use cases on Deduped" height="20"></a>

A static image, no tracking script. About the badge

Questions about Open Shell

What is Open Shell?

OpenShell is NVIDIA's open-source runtime for running autonomous AI agents safely. Each agent gets an isolated sandbox where kernel controls confine file access and system calls, every network connection passes a policy check, and real credentials are only added to requests bound for approved endpoints. Policy changes are checked with formal verification and risky new access waits for human review. It is an open-source project on GitHub: https://github.com/NVIDIA/OpenShell.

Who is Open Shell for?

Developers and platform teams who want coding agents to install packages, call APIs and use credentials without giving them the whole machine.

What can I do with Open Shell?

3 use cases so far, for example: Run a coding agent in a locked-down sandbox; Let agents use API keys they can never see; Teach your coding agent to drive Open Shell. Each has steps and a copy-paste agent prompt.

What do I need to run Open Shell?

Linux, macOS on Apple Silicon, or Windows with WSL 2 (experimental); Docker, Podman, or host virtualization; An agent to run inside the sandbox; the default image is minimal Ubuntu with no agent installed; For Kubernetes: Helm and a CNI that enforces NetworkPolicy.

What are alternatives to Open Shell?

Other hot AI repos in Software and DevOps: Deep Infer, Harness2, Orca, Opencode, Graphify.

Is Open Shell safe to run?

Deduped does not own, audit or vouch for it. Read the README and license, run it on your own systems first and review every command before you run it.