"A tool or assistant to explore and debug my codebase" matching MCP servers:

Evidence-backed, read-only codebase diagnostics for coding agents: secrets, dependency, RLS, and source-impact audits with a verify loop and PR review verdicts. Never writes files or uses the network by default.

Trust layer for 412 official Malaysian public datasets, with freshness, licences, and a public read-only MCP server. No API key.

Rust-based MCP server that lets AI agents natively control and debug Chromium browsers via the Chrome DevTools Protocol (CDP). Single pre-compiled binary, installable with npx @raultov/chrome-debug-mcp, Homebrew or one shell command — no Rust toolchain.

MCP router that fronts your other MCP servers behind a few small generic tools, so their tool definitions stay out of the context window and are discovered on demand.

Enables AI agents to inspect and modify live UI layouts (web DOM or Android Compose) by selecting and dragging elements on the real screen, providing precise element details like source location, anchors, and tweaks for code changes.

Provides comprehensive development tools for FastMCP projects including documentation access, NPM package version management, and TypeScript type definitions retrieval. Enables developers to fetch FastMCP documentation, analyze NPM packages, and access MCP architecture information through natural language.

Codebase Q&A agent with git-history awareness. Explore a repo with list_files, read_file, and search, plus git log, blame, and diff and a repo map to answer "why did this change?" questions. Runs locally as an MCP server for Claude Code and other MCP clients.

Local-first MCP runtime for coding agents with bounded workspace access, content-preconditioned writes, AST search, LSP navigation and diagnostics, and Git status/diff. Designed to give coding agents useful repository access without treating unrestricted shell access as the default integration boundary.

Enables local-first codebase intelligence, allowing chat, search, and audit operations on any repository with file:line citations, and supports offline deterministic modes without an LLM.

Record debug sessions, commands, failed attempts, and successful fixes in local SQLite storage.

Enables AI IDEs to chat with, search, explain, refactor, and audit a local codebase without uploading source code, returning answers backed by file-and-line source citations, confidence scores, and severity labels, plus audit and board-ready strategic reports. Runs locally and standalone as an MCP server compatible with Claude, Cursor, Windsurf, and DeepSeek Harness.

Run free live audience Q&A and polls from your agent: start an AskOne room, add and launch a poll, approve or hide questions and mark them answered, close the room, then read the questions as a FAQ draft and get the results. Free AI moderation. Uses an organization API token from askone.org

Enables agents to convert Markdown into WhatsApp-compatible formatting, including styled text, lists, and tables that fit a phone's monospace width.

Enables AI assistants to convert CSS-animated SVGs into videos (MP4/WebM/MKV/MOV) and animated images (aPNG/GIF) with frame-accurate rendering and transparency support.

22 consolidated MCP tools for email, calendar, contacts, folders, rules, categories, and settings with safety controls (dry-run preview, rate limiting, recipient allowlists) and MCP annotations on every tool. Works with personal Outlook.com and work/school Microsoft 365 accounts.

Audits MCP tool descriptions for quality and reliability, scoring them 0-100, detecting smells, and providing rewritten versions for better agent accuracy.

Graph-based tool retrieval for LLM agents. Builds a tool graph from OpenAPI/MCP specs and retrieves multi-step workflows via hybrid search (BM25 + graph traversal + embedding), recovering accuracy from 12% to 82% with 79% fewer tokens. Also works as an MCP Proxy to aggregate multiple servers behind 3 meta-tools.

Tell your AI agent what you want a model to do. It searches Hugging Face for each part of the job and tells you which parts already exist there: which are an exact fit, which to fine-tune, which to train, and which are just code. Every model it names is real, current, and sized for your hardware.

Enables local Laya-MLX typed decisions (classification, scoring, risk routing, and yes/no) for coding agents like Codex, Claude Code, DeepSeek Harness, and Pi via MCP, running entirely on Apple Silicon Macs.

A deterministic MCP server that executes 10-step topology pipelines on Cloudflare Workers, providing structured progress snapshots and adaptive JSON templates for AI agents.