NVIDIA/OpenShell-Community

NVIDIA/OpenShell-Community

Description: OpenShell is the safe, private runtime for autonomous AI agents.

Language: Dockerfile

License: Apache-2.0

Stars: 156

Forks: 52

Open issues: 16

Created: 2026-03-01T03:52:47Z

Pushed: 2026-05-29T16:15:11Z

Default branch: main

Fork: no

Archived: no

README:

OpenShell Community

OpenShell is the runtime environment for autonomous agents -- the infrastructure where they live, work, and verify. It provides a programmable factory where agents can generate synthetic data to fix edge cases and safely iterate through thousands of failures in isolated sandboxes. The core engine includes the sandbox runtime, policy engine, gateway (with k3s harness), privacy router, and CLI.

This repo is the community ecosystem around OpenShell -- a hub for contributed skills, sandbox images, and integrations that extend its capabilities. For the core engine, docs, and published artifacts (PyPI, containers, binaries), see the OpenShell repo.

> Alpha software — single-player mode. OpenShell is proof-of-life: one developer, one environment, one gateway. We are building toward multi-tenant enterprise deployments, but the starting point is getting your own environment up and running. Expect rough edges. Bring your agent.

What's Here

| Directory | Description | | ------------ | --------------------------------------------------------------------------------- | | sandboxes/ | Pre-built sandbox images for domain-specific workloads (each with its own skills) |

Sandboxes

| Sandbox | Description | | ----------------------- | ------------------------------------------------------------ | | sandboxes/base/ | Foundational image with system tools, users, and dev environment | | sandboxes/droid/ | Android automation and mobile testing workflows | | sandboxes/gemini/ | Gemini CLI workflows | | sandboxes/nvidia-gpu/ | GPU-enabled VM sandbox image with NVIDIA userspace tooling | | sandboxes/ollama/ | Ollama for local and cloud LLMs with Claude Code, Codex, OpenCode pre-installed | | sandboxes/pi/ | Pi pre-installed | | sandboxes/sdg/ | Synthetic data generation workflows |

Getting Started

Prerequisites

• OpenShell CLI installed (uv pip install openshell)
• Docker or a compatible container runtime
• NVIDIA GPU with appropriate drivers (for GPU-accelerated images)

Using Sandboxes

openshell sandbox create --from ollama

The --from flag accepts any sandbox defined under sandboxes/ (e.g., ollama, sdg), a local path, or a container image reference.

Ollama Sandbox

The Ollama sandbox provides Ollama for running local LLMs and routing to cloud models, with Claude Code and Codex pre-installed.

Quick start:

openshell sandbox create --from ollama

curl http://127.0.0.1:11434/api/tags

See the [Ollama sandbox README](sandboxes/ollama/README.md) for full details.

Contributing

See [CONTRIBUTING.md](CONTRIBUTING.md).

Security

See [SECURITY.md](SECURITY.md). Do not file public issues for security vulnerabilities.

License

This project is licensed under the Apache 2.0 License -- see the [LICENSE](LICENSE) file for details.