"author:delonsp" matching MCP servers:

Enables the analysis of massive datasets by storing data in external variables via a sandboxed Python REPL instead of the model's context window. This allows users to process files larger than 100MB without polluting the context or hitting token limits.