使用说明
🛠️ AI开发工具包
一系列基于人工智能的开发工具——MCP服务器和从头开始构建的人工智能代理。
📌 这是什么?
此仓库包含在学习创建MCP服务器和AI代理时构建的动手AI工程项目。每个项目都有详细的学习笔记,解释了 什么、为什么以及如何.
建造人: 全栈(MERN)开发人员过渡到人工智能工程。
📂 项目结构
ai-dev-toolkit/
├── mcp-server/ # MCP Server — Developer Toolkit
│ ├── app.py # Shared FastMCP instance
│ ├── server.py # Entry point (stdio transport)
│ ├── server_sse.py # Entry point (SSE/HTTP transport)
│ ├── validators.py # Input validation & security
│ ├── tools/
│ │ ├── dev_tools.py # 9 core tools (Phase 1)
│ │ ├── api_tools.py # 3 async API tools (Phase 3)
│ │ ├── db_tools.py # 4 database CRUD tools (Phase 4)
│ │ └── advanced_tools.py # 3 advanced tools (Phase 6)
│ ├── resources/
│ │ └── system_resources.py # 6 system resources (Phase 2)
│ ├── prompts/
│ │ └── code_prompts.py # 3 code prompts (Phase 2)
│ ├── db/
│ │ └── database.py # SQLite setup & connection
│ ├── tests/
│ │ └── test_validators.py # 37 unit tests (Phase 7)
│ ├── TUTORIAL.md # Phase 1 build tutorial
│ ├── PHASE2_TUTORIAL.md # Resources & Prompts tutorial
│ ├── PHASE3_TUTORIAL.md # Async API tools tutorial
│ ├── PHASE4_TUTORIAL.md # Database tools tutorial
│ ├── PHASE5_TUTORIAL.md # SSE transport tutorial
│ ├── PHASE6_TUTORIAL.md # Advanced patterns tutorial
│ ├── PHASE7_TUTORIAL.md # Production patterns tutorial
│ └── requirements.txt # Python dependencies
├── docs/
│ ├── mcp-learning-roadmap.md # Learning roadmap
│ ├── mcp-mastery-notes.md # Concept reference
│ └── TUTORIAL.md # Getting started guide
├── AI_ENGINEER_ROADMAP.md # AI Engineering learning path
├── requirements.txt
├── .gitignore
└── README.md
🔧 MCP服务器——开发人员工具包
一种模型上下文协议(MCP)服务器 19工具, 6资源,以及 3个提示 AI助手。
工具(19)
| 类别 | 工具 | 文件 |
|---|---|---|
| 核心开发工具 | greet, generate_uuid, format_json, base64_encode, base64_decode, hash_text, word_count, timestamp_convert, regex_test |
tools/dev_tools.py |
| API工具 | search_github, get_weather, fetch_url |
tools/api_tools.py |
| 数据库工具 | create_note, list_notes, search_notes, delete_note |
tools/db_tools.py |
| 先进工具 | system_health_check, smart_note_search, bulk_create_notes |
tools/advanced_tools.py |
资源(6)
system://info, system://env, system://packages, project://structure, config://server, docs://readme
提示(3)
code_review, explain_code, debug_error
快速开始
# Clone the repo
git clone https://github.com/ravirajput10/ai-dev-toolkit.git
cd ai-dev-toolkit/mcp-server
# Create virtual environment
python -m venv venv
# Activate (Windows PowerShell)
.\venv\Scripts\Activate.ps1
# Activate (Mac/Linux)
source venv/bin/activate
# Install dependencies
pip install -r requirements.txt
# Test with MCP Inspector
mcp dev server.py
以远程服务器(SSE)运行
python server_sse.py # Default: localhost:8000
python server_sse.py --host 0.0.0.0 --port 9000 # Custom
运行单元测试
python tests\test_validators.py -v # 37 tests
连接到AI客户端
{
"mcpServers": {
"dev-toolkit": {
"command": "python",
"args": ["path/to/ai-dev-toolkit/mcp-server/server.py"]
}
}
}
🧠 学习笔记
每个阶段都包括详细的教程,包括:
- 映射到Express/Node.js类比的概念解释
- 逐步构建指南
- 修订表和关键模式
📖 从...开始 TUTORIAL.md → 然后是第2-7阶段教程。
🛠️ 技术栈
| 技术 | 目的 |
|---|---|
| Python 3.10+ | 核心语言 |
MCP-SDK(FastMCP) |
MCP服务器框架 |
httpx |
异步HTTP客户端(API工具) |
| SQLite | 数据库(笔记管理器) |
unittest |
单元测试 |
📄 许可证
MIT许可证——您可以自由使用、修改和学习此代码。
