🧩 Jcodemunch MCP: AST-Based Code Retrieval

This project addresses a common pain point in AI-assisted development: flooding the context window with irrelevant code. It is an MCP (Model Context Protocol) server designed for precise, symbol-level code retrieval from GitHub repositories using tree-sitter abstract syntax trees (AST). The interesting part is its structural approach to code search. Instead of relying on full-text search or basic vector embeddings, it uses AST to understand the code's architecture. When an AI client like Cursor or Claude Code needs context on a specific function or class, the server fetches only the relevant structural snippet rather than the entire file, significantly reducing token usage. For developers who frequently point language models at large codebases, this is a highly practical utility. It effectively bridges traditional static code analysis with modern LLM context protocols to make code exploration much more efficient.