mcp_code_executor
The MCP Code Executor is an MCP server that facilitates the execution of Python code for language models, enabling complex and large-scale code generation beyond token limits with support for various Python environments.
What environments are supported by the MCP Code Executor?
The MCP Code Executor supports Conda, standard virtualenv, and UV virtualenv environments.
How does the server handle large code blocks?
The server supports incremental code generation, allowing LLMs to initialize a file, append code in parts, and execute the complete code without exceeding token limits.
Can the environment configuration be changed dynamically?
Yes, the environment configuration can be dynamically changed at runtime using the 'configure_environment' tool.
Is backward compatibility maintained?
Yes, the package maintains backward compatibility with earlier versions, especially for users who specified a Conda environment.
What is the purpose of the 'initialize_code_file' tool?
The 'initialize_code_file' tool is used to create a new Python file with initial content, serving as the first step for longer code that may exceed token limits.