This lets Claude execute Python, JavaScript, and shell commands in a sandboxed environment with timeout protection and output capture. You get three execution tools that run code snippets safely without touching your local system directly. Reach for it when you need Claude to prototype scripts, test code blocks, or run calculations that benefit from actual execution rather than simulation. The sandbox includes safety guards to prevent resource abuse and captures both stdout and stderr. Built by MEOK AI Labs with EU AI Act compliance hooks baked in, though the core functionality is straightforward code execution wrapped in MCP protocol calls.
claude mcp add --transport stdio csoai-org-code-executor-mcp -- uvx code-executor-mcp