Give MCP agents disposable code execution sandboxes with OpenSandbox MCP

Expose sandbox creation, command execution, and file operations to MCP-capable agents so they can run untrusted code inside disposable Docker or Kubernetes-backed runtimes.

Prerequisites

Docker for local runtime or Kubernetes for cluster runtime, plus an MCP-capable client such as Claude Code or Cursor.

Installation

Use the upstream install or setup path that matches your environment:

• pip install opensandbox
• npm install @alibaba-group/opensandbox
• pip install opensandbox-cli
• uv tool install opensandbox-cli

Requirements and caveats from upstream:

• OpenSandbox is a general-purpose sandbox platform for AI applications, offering multi-language SDKs, unified sandbox APIs, and Docker/Kubernetes runtimes for scenarios like Coding Agents, GUI Agents, Agent Evaluat...
• Multi-language SDKs: Provides sandbox SDKs in Python, Java/Kotlin, JavaScript/TypeScript, C#/.NET, Go.
• Sandbox Runtime: Built-in lifecycle management supporting Docker and high-performance Kubernetes runtime, enabling both local runs and large-scale distributed scheduling.

Basic usage or getting-started notes:

• OpenSandbox also provides osb, a terminal CLI for the common sandbox workflow: create sandboxes, run commands, move files, inspect diagnostics, and manage runtime egress policy.

• bash

• or

• Source: https://github.com/alibaba/OpenSandbox

• Extracted from upstream docs: https://raw.githubusercontent.com/alibaba/OpenSandbox/HEAD/README.md

Documentation

• https://open-sandbox.ai

Source

• Agent Skill Exchange