Emilien Macchi
07b7a74b62
The provisioning-deadline scan takes each provisioning sandbox's local S(global) S(workspace) X(sandbox) keys in ascending key order and waited for them with no deadline. When a sandbox's key sorts before its workspace's key, the scan held X(sandbox) while it waited behind a provider or profile writer, so start, stop and delete for that sandbox waited for the writer too. In the other order, the scan held S(workspace) while it waited for a busy sandbox, and provider writers on that workspace queued behind it. Either way the rest of the pass waited as well. Each pass now has a 500 ms budget for waiting on these keys. A sandbox whose keys stay busy is skipped and picked up on the next pass a second later. Once the budget is spent, the rest of the pass only takes keys that are free when it reaches them, so a busy workspace or global key costs one wait per pass rather than one per sandbox. The local acquisition is cancel-safe, so giving up releases the keys it already took. Skips log at debug and count in the new openshell_server_provisioning_deadline_skips_total counter, labeled by whether the sandbox was already past its deadline. A steady rate of delayed expiries can then be alerted on without a warning every pass. The gateway metrics page and the cluster debugging skill describe it. The test that pinned the old wait is replaced by tests for both key orders: a lifecycle lock or a provider writer proceeds and the scan finishes while the other side holds its key, the skipped sandbox is untouched and expires once the writer releases, and a skip doesn't end the pass for other sandboxes. Another test checks that sandboxes behind one busy workspace share a single wait. Part of #4307 Closes #4547 Signed-off-by: Emilien Macchi <emacchi@redhat.com>
Important
New in OpenShell 0.1.x: a stable release cadence, new isolation primitives, an expanded extension surface, and new APIs. Read the 0.1.0 upgrade guide.
OpenShell is the safe, private runtime for fleets of autonomous AI agents. Agents are most useful when they can read files, install packages, call APIs, and use credentials. OpenShell gives them that capability without giving them unrestricted access to your data, secrets, or network. You declare what each agent can touch in a policy, and OpenShell enforces it.
How It Works
OpenShell governs what agents can do in two ways: it instruments the kernel to enforce policy on every file access, system call, and network connection at runtime, and it uses formal verification to check what a policy change would allow before it is applied.
- Kernel-level enforcement. Each agent runs in an isolated sandbox. Kernel controls confine which files it can access and which system calls it can make, and every network connection passes through a policy check before it leaves the sandbox. Agents never see real credentials; OpenShell adds them only to requests bound for approved endpoints.
- Formally verified policy changes. Before a policy change is approved, OpenShell uses formal verification to flag risky new access it would grant, such as reaching a new host with credentials or calling a new API method, so those changes wait for human review.
See Architecture for how the gateway, supervisor, and sandbox fit together.
Quickstart
You need Linux, macOS on Apple Silicon, or Windows with WSL 2 (experimental), plus Docker, Podman, or host virtualization. See the Support Matrix for details.
curl -LsSf https://raw.githubusercontent.com/NVIDIA/OpenShell/main/install.sh | sh
openshell sandbox create --name demo
The installer sets up the CLI and a local gateway. The default sandbox image is minimal Ubuntu with no agent installed. To run a real agent, follow Run Your First Agent: it runs OpenCode against a free OpenRouter model and shows how to approve new access as the agent needs it.
Explore Further
- Sandboxes: images, runtimes, GPUs, and lifecycle.
- Policies: filesystem, network, and process rules, with the advisor and prover for reviewing changes.
- Providers: credentials that work only at approved endpoints, including inference.
- Gateways: the control plane for sandboxes, policy, and access.
- Kubernetes: deploy the gateway with Helm. Your CNI must enforce
NetworkPolicy. - Extensibility: middleware, interceptors, and compute drivers.
- Tutorials: step-by-step policy and agent walkthroughs.
- Prerelease and development builds: try an upcoming release or the latest commit on
main.
Agent Skills
Install the public OpenShell skills for your coding agent:
npx skills add NVIDIA/OpenShell
The skills teach your agent to drive the OpenShell CLI, write sandbox policies, and debug gateways and inference routing. They live in skills/ and work without an OpenShell source checkout.
SDKs
SDKs connect applications to an OpenShell gateway. They do not install the CLI. Use the same OpenShell release for the SDK and the gateway when possible.
| Language | Install | Docs |
|---|---|---|
| Python | uv add openshell |
README |
| TypeScript | npm install @nvidia/openshell-sdk (GitHub Packages) |
README |
| Go | go get github.com/NVIDIA/OpenShell/sdk/go@latest |
README |
| Rust | cargo add openshell-sdk --git https://github.com/NVIDIA/OpenShell --tag <release-tag> |
Installation and usage |
Community
- Questions and discussion: GitHub Discussions
- Bug reports and feature requests: GitHub Issues, using the issue templates
- Security vulnerabilities: follow SECURITY.md. Do not open a GitHub issue.
- Roadmap: OpenShell Roadmap and the RFC board
- Try it in the cloud: Brev Launchable
OpenShell is built agent-first: it is developed with the same agent-driven workflows it enables. See CONTRIBUTING.md for development setup and the contribution workflow, and AGENTS.md for repository coding conventions.
Telemetry
OpenShell collects anonymous telemetry, limited to operational categories and counts, to help improve the project. It does not collect sandbox names, hostnames, file paths, prompts, credentials, provider or model names, or user content. To disable it, set OPENSHELL_TELEMETRY_ENABLED=false on the gateway, or server.telemetryEnabled=false for Helm installs. You can also compile telemetry out entirely. See Telemetry for details and the community telemetry reports for published usage trends.
Notice and Disclaimer
This software automatically retrieves, accesses or interacts with external materials. Those retrieved materials are not distributed with this software and are governed solely by separate terms, conditions and licenses. You are solely responsible for finding, reviewing and complying with all applicable terms, conditions, and licenses, and for verifying the security, integrity and suitability of any retrieved materials for your specific use case. This software is provided "AS IS", without warranty of any kind. The author makes no representations or warranties regarding any retrieved materials, and assumes no liability for any losses, damages, liabilities or legal consequences from your use or inability to use this software or any retrieved materials. Use this software and the retrieved materials at your own risk.
License
This project is licensed under the Apache License 2.0.