MCP Code Sandbox provides secure, isolated environments for executing Python code through multiple backend options including E2B and Firecracker microVMs. Built with a modular architecture, it offers tools for sandbox creation, code execution, file operations, package installation, and data visualization through chart generation. The implementation features an abstract interpreter interface that allows seamless switching between execution backends, robust error handling, and resource cleanup mechanisms. This server is particularly valuable for AI assistants that need to run untrusted code, test programming solutions, generate data visualizations, or interact with files in a controlled environment without security risks.
暂无评论. 成为第一个评论的人！
登录以参与讨论
Create a new sandbox environment.
Close and clean up a sandbox.
Check status of sandboxes.
Run Python code in a sandbox.
Install a Python package.
All-in-one tool that creates a sandbox, runs code, and cleans up.
List files in the sandbox.
Read the contents of a file.
Write content to a file.
Upload a file to the sandbox.
