我創建了一個讓 AI 能進行運行時程式碼偵錯(斷點、逐行執行等)的 MCP

我創建了一個讓 AI 能進行運行時程式碼偵錯(斷點、逐行執行等)的 MCP

Hacker News·

本文介紹了 AI Debugger Backend (AIDB),這是一個為 AI 代理設計的程式化、語言無關的偵錯介面。它旨在將 Debug Adapter Protocol (DAP) 生態系統帶入 AI 領域,實現 AI 驅動的程式碼執行中的斷點和逐行執行等功能。

Navigation Menu

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

To see all available qualifiers, see our documentation.

AI Debugger Backend: a programatic, language-agnostic debugging interface for AI agents.

License

Uh oh!

There was an error while loading. Please reload this page.

ai-debugger-inc/aidb

Folders and files

Latest commit

History

Repository files navigation

AI Debugger

Image

Image

Image

Image

Image

AI-Powered Debugging for Every Language

AI Debugger (AIDB) brings the proven Debug Adapter Protocol (DAP) ecosystem to
AI agents through a standardized Model Context Protocol (MCP) interface. Debug
Python, JavaScript, TypeScript, and Java programs using the same battle-tested
adapters that power VS Code—no IDE required, no heavyweight dependencies, just
powerful debugging at your AI assistant's fingertips.

Read the Docs | Join
Discord | Star on
GitHub

Quick Install

Get started with Python debugging in under 60 seconds:

Add to your MCP client settings (Claude Code, Cline, Cursor, etc.):

Ask your AI assistant:

"Initialize debugging for Python. Debug app.py with a breakpoint at line
25."

JavaScript/Java? Visit the
docs for
multi-language setup.

Why AI Debugger?

Standalone & Zero Heavy Dependencies

No VS Code required. No heavyweight IDEs. Just install with pip and you're
debugging––works on macOS, Linux, and Windows (WSL supported).

The core Python dependencies are lightweight and minimal:

Debug adapters are built during the release pipeline and are published as
release artifacts. Once the ai-debugger-inc package is installed, your agent
will use the download tool to fetch the appropriate adapter binaries
automatically on first run.

Multi-Language from Day One

Debug Python, JavaScript, TypeScript, and Java with a single MCP server. AIDB is
designed to support all DAP-compatible adapters, with more languages coming.

Built on the DAP Standard

AIDB uses the same Debug Adapter Protocol that powers VS Code debugging. We
integrate with proven, open-source debug adapters:

This means you get reliable, well-maintained debugging that "just works" with
established patterns developers already trust.

VS Code Integration (Without VS Code)

Already have complex debug configurations in launch.json? AIDB can use them
directly—making sophisticated debugging setups portable and shareable across
teams without requiring VS Code installations.

Advanced Debugging Features

Future-Ready Architecture

AIDB is built for where AI-assisted development is heading:

How It Works

The Bridge Between AI and Proven Tools

AI Debugger acts as a translation layer, exposing the mature Debug Adapter
Protocol ecosystem to AI agents through a clean, agent-optimized MCP interface.
Your AI assistant gets powerful debugging capabilities, and you get the
reliability of debug adapters used by millions of developers worldwide.

Learn more about the architecture

Supported Languages

Platforms: macOS, Linux, Windows (WSL recommended; native support in progress)

Coming Soon: Built to support all DAP-compatible adapters––AIDB is designed
to become the debugging standard for AI systems across every popular language
and framework.

Documentation

Getting Started

Technical Reference

Architecture & Design

Development Setup

Prerequisites: Python 3.10+, Docker

Initial setup:

Common commands:

Project Structure

For architecture details and implementation guidance, see the Developer
Guide.

Robust Testing & Releases

AIDB is built with a comprehensive CI/CD pipeline:

We catch issues early and ship features confidently, ensuring the debugging
experience you depend on stays reliable.

Our entire CI/CD release pipeline executes start to finish in under 15
minutes––a target we plan to maintain.

Our Vision

Becoming the debugging standard in the MCP tools space.

As AI agents become more capable, they need debugging tools designed for their
workflows—not adapted from human-centric IDEs. AIDB provides a unified,
language-agnostic approach to debug any program with any AI agent through the
proven MCP standard.

We're building the future of AI-assisted debugging, one DAP adapter at a time.

Contributing

We welcome contributions! See our Contributing Guide to get
started.

Community & Support

License

AI Debugger is licensed under the Apache 2.0 License. See LICENSE for
details.

Ready to bring debugging to your AI assistant?

Get Started
| Read the Docs | Join
Discord

About

AI Debugger Backend: a programatic, language-agnostic debugging interface for AI agents.

Resources

License

Code of conduct

Contributing

Security policy

Uh oh!

There was an error while loading. Please reload this page.

Stars

Watchers

Forks

Releases

  2

Contributors

  3

Image

Image

Image

Languages

Footer

Footer navigation

Hacker News

相關文章

  1. 瀏覽器開發者工具MCP:AI代理用於網頁程式碼的自我測試與除錯

    3 個月前

  2. Show HN:ArchitectGBT MCP - AI 輔助開發的智慧模型選取

    3 個月前

  3. 適用於 AI 編碼代理的語言伺服器協定 (LSP)

    4 個月前

  4. 開發者基礎:使用 AI 建構軟體所需的最低知識

    3 個月前

  5. Show HN:Apex Agent – 透過 MCP 將瀏覽器連接至 AI

    3 個月前