Azure DevOps MCP Server
Integrates Azure DevOps with GitHub Copilot Chat, enabling natural language queries of sprint work items, ticket summaries, and status tracking directly from VS Code without switching contexts.
README
Azure DevOps MCP Server
A Model Context Protocol (MCP) server that integrates Azure DevOps with GitHub Copilot Chat, allowing you to query work items naturally through conversation
🚀 What is MCP?
Model Context Protocol (MCP) is an open standard that connects AI assistants like GitHub Copilot to external data sources. This tool implements an MCP server that:
- 🔌 Connects GitHub Copilot to your Azure DevOps project
- 💬 Enables natural language queries - just ask Copilot about your tickets
- ⚡ Provides instant responses from cached data
- 🔒 Keeps your data local - no third-party services involved
Instead of switching between VS Code and Azure DevOps web portal, you can ask Copilot directly:
- "Show my sprint tickets"
- "Summarize ticket 13080"
- "What tickets are in ACC?"
🎯 What This Does
This tool helps you:
- Instantly view all User Stories where you have verification tasks assigned
- Query cached data without API rate limits or permission prompts
- Summarize tickets with full details including comments and descriptions
- Track your sprint work efficiently
- Stay in your IDE - no context switching needed
✨ Key Features
- 🔌 MCP Protocol Implementation - Native integration with GitHub Copilot via Model Context Protocol
- 🚀 Cache-based queries - Instant responses from local cache
- 🔍 Smart filtering - Automatically finds User Stories with your query
- 💬 Full ticket details - Descriptions, comments, state, assignees, tags
- 📊 Sprint tracking - Focused on current sprint items only
- 🤖 Natural language interface - Ask Copilot in plain English, no special syntax needed
🛠️ Setup
1. Prerequisites
- Node.js installed
- Azure DevOps personal access token
2. Configuration
Create a .env file with your credentials:
AZURE_DEVOPS_ORG_URL=https://your-org.visualstudio.com
AZURE_DEVOPS_TOKEN=your-personal-access-token
AZURE_DEVOPS_PROJECT=your-project-name
AZURE_DEVOPS_USER_EMAIL=your.email@company.com
3. Install Dependencies
npm install
npm run build
📖 How to Use
Step 1: Refresh Cache (Once per day or when needed)
npm run refresh
This fetches all User Stories from the current sprint where you have a "Verify" task assigned. Takes ~10-30 seconds depending on the number of stories.
Output example:
Found 12 user stories with your verify tasks
✓ Cache updated successfully!
Work items: 12
Last updated: 2025-11-17T07:33:36.435Z
Step 2: Query Your Tickets (Instant, no prompts!)
After caching, you can query instantly using GitHub Copilot Chat in VS Code:
In Copilot Chat, ask:
show all my sprint user storiessummarize ticket 13080show all tickets in status "In ACC"show tickets with tag "HIGHPRIO"
Or use the CLI:
# List all your verify stories
node cli.js my-verify
# Get details of a specific ticket
node cli.js item 13080
# Query cached data
npm run query
Query Cache Commands
npm run query
Available commands:
list- List all cached user storiesitem <id>- Show details for a specific ticketby-state <state>- Filter by state (Active, "In ACC", New, etc.)info- Show cache information
🎨 MCP Server Integration with GitHub Copilot
What Makes This Special?
This tool implements the Model Context Protocol (MCP), which means:
- 🔌 Direct Integration - Copilot can access your Azure DevOps data as if it's a native feature
- 💬 Natural Conversation - No special commands or syntax needed
- ⚡ Real-time Context - Copilot knows about your tickets while you code
- 🔒 Secure & Local - Your data stays on your machine, accessed via your personal token
How It Works
┌─────────────────┐
│ GitHub Copilot │ "Show my sprint tickets"
│ in VS Code │
└────────┬────────┘
│ MCP Protocol
▼
┌─────────────────┐
│ MCP Server │ Queries local cache
│ (This Tool) │
└────────┬────────┘
│
▼
┌─────────────────┐
│ Cached Data │ Your sprint tickets
│ work-items.json│ (refreshed from Azure DevOps)
└─────────────────┘
### MCP Configuration
The MCP server is configured in `.vscode/settings.json`:
```json
{
"github.copilot.chat.codeGeneration.instructions": [
{
"file": "azure-devops-mcp-server"
}
],
"mcp.servers": {
"azure-devops": {
"command": "node",
"args": ["dist/index.js"],
"cwd": "${workspaceFolder}/azure-devops-mcp-server"
}
}
}
This tells VS Code to run the MCP server when Copilot needs Azure DevOps data.
📁 Project Structure
azure-devops-mcp-server/
├── src/
│ ├── azure-devops-client.ts # Azure DevOps API client
│ └── index.ts # MCP server implementation
├── cache/
│ └── work-items.json # Cached sprint data
├── cli.js # Command-line interface
├── refresh-cache.js # Cache refresh script
├── query-cache.js # Cache query script
├── .env # Your credentials (gitignored)
└── package.json # Project configuration
🔄 Typical Workflow
- Morning: Run
npm run refreshto get today's sprint data - During work: Ask Copilot about tickets - instant responses from cache
- Need updates? Run
npm run refreshagain
🎯 Why This Matters
Before this tool:
- Every query required clicking "Allow" button
- Interrupts your flow
- Slow responses
With this tool:
- Cache once, query unlimited times
- No permission prompts
- Instant responses in Copilot Chat
🚀 Quick Start for Team Members
- Clone/copy this folder to your machine
- Create
.envfile with your Azure DevOps credentials - Install dependencies:
npm install && npm run build - Refresh cache:
npm run refresh - Configure VS Code to enable MCP server (see MCP Configuration above)
- Start chatting with Copilot about your tickets!
💡 Pro Tip: Once the MCP server is running, Copilot becomes your Azure DevOps assistant. Just ask questions naturally!
💡 Tips
- Refresh cache at the start of your workday
- Cache includes all comments and descriptions
- Works offline once cached
- Cache file is human-readable JSON (check
cache/work-items.json)
🔧 Troubleshooting
"Error: azureClient.getUserStoriesWithMyVerifyTasks is not a function"
- Run
npm run buildto recompile TypeScript
"Found 0 user stories"
- Check your email in
.envmatches Azure DevOps - Verify you have "Verify" tasks assigned in current sprint
Cache is outdated
- Run
npm run refreshto update
📝 License
ISC
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。