Codebase analysis MCP server for AI agents: complexity metrics, size thresholds, prop drilling, and code smell detection built on tree-sitter and tokei.
It gives the agent 8 read-only tools: line counts, cognitive complexity/nesting/parameter counts, an oversized-file/function checker against configurable thresholds, doc coverage, code smell detection (TODO/FIXME/HACK in comments and unsafe casts), and prop-drilling detection. It's tree-sitter + tokei + fast-glob under the hood; nothing is custom-parsed. Tested on macOS with Claude Code, but should work with Cursor, VS Code, Windsurf, Gemini CLI, Codex, or anything else that speaks MCP.
TS/JS, Python, Go, Rust, Java, C/C++, and Ruby.
I've actually been sitting on this for a while, because it's mostly vibe-coded. But I've used it and kept refining and fixing it for months, and it seems to work. Plus my colleague finally pushed me into polishing the repo and making it public.
Real-time code health TUI with Copilot integration, but "chat with codebase" category is crowded.
MCP wrapper around codebase diagramming when Mermaid and PlantUML already exist.
Multi-protocol agent deployment, but MCP and orchestration are solved by Anthropic, LangChain.
MCP query server cuts codebase context by 87% with zero dependencies, measured.
Spec-driven PR automation with live agent verification, not chat babysitting.
Indexes the Linux kernel in 80 seconds without AST parsers, where tree-sitter tools fail.
