Execute terminal commands securely within a restricted workspace, enabling LLM clients to perform local machine operations.
This Model Context Protocol (MCP) server empowers large language model clients, such as Claude, to interact with your local machine's terminal environment safely. It provides a robust and secure mechanism for executing shell commands, strictly confined to a dedicated workspace directory. This setup prevents unintended system-wide modifications, making it an invaluable tool for agent-driven workflows requiring file operations, scripting, and automated tasks within a controlled sandbox.
Características Principales
01Runs shell commands via MCP tool interface
02Restricts execution to a dedicated workspace directory
03Prevents accidental execution outside the workspace
04Lightweight and easily extensible
05Works across Windows, macOS, and Linux operating systems
060 GitHub stars
Casos de Uso
01Automating file operations in agent-driven workflows
02Executing scripts as part of LLM-orchestrated tasks
03Enabling secure interaction with the local filesystem for AI agents
