Exec Sandbox

Overview

Exec Sandbox is a Model Context Protocol server that provides a single exec tool for running Python code in a network-isolated, locked-down sandbox. Any MCP-compatible AI assistant can call the tool directly, with the model deciding when to execute code during conversations or agent runs.

The server runs locally via the Docker image docker.io/idconstruct/mcp-exec:v0.5.4, communicating over stdio. It requires an API key for authentication. Version v0.5.4 is the latest release, published under the MIT license with source code available at inhuman/mcp-exec on GitHub.

Exec Sandbox is categorized as a Development tool on mcp.site and works with MCP clients including Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, Zed, and others. If you maintain Exec Sandbox, claim this listing to verify ownership and keep information current.

Installation

// claude_desktop_config.json
{
  "mcpServers": {
    "exec_sandbox": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "docker.io/idconstruct/mcp-exec:v0.5.4"
      ]
    }
  }
}