Azure PR MCP Server
An MCP server that enables reviewing Azure DevOps Pull Request comments using the Azure CLI, with features like comment retrieval and status filtering.
README
Azure PR MCP Server
A Model Context Protocol (MCP) server for reviewing Azure DevOps Pull Request comments using the Azure CLI.
🔒 Readonly.
🤖 Made by the AI for the AI.
Features
- ✅ Azure CLI Validation - Automatically checks for Azure CLI availability
- 🔐 Authentication Handling - Verifies Azure CLI authentication status
- 💬 PR Comment Retrieval - Fetches PR comments with full metadata
- 🔍 Status Filtering - Filter comments by status (active, fixed, closed, etc.)
- 🧪 Unit Tests - Comprehensive test coverage using Vitest
Prerequisites
- Node.js >= 18
- pnpm >= 9.0.0
- Azure CLI - Installation guide
- Azure DevOps Organization - Access to an Azure DevOps organization with pull requests
Installation
Install dependencies
pnpm install
Build the project
pnpm build
Azure CLI Setup
-
Install Azure CLI (if not already installed):
- macOS:
brew install azure-cli - Windows: Download from Microsoft Docs
- Linux: Follow Linux installation guide
- macOS:
-
Login to Azure:
az login -
The Azure DevOps extension will be automatically installed when you first use the server.
Usage
Running the MCP Server
The server runs on stdio and is designed to be used with MCP clients:
node dist/index.js
Available Tools
1. check_azure_cli
Checks if Azure CLI is installed and the user is authenticated.
Input: None
Output:
{
"cli_installed": true,
"cli_version": "2.50.0",
"authenticated": true,
"account": "user@example.com",
"devops_extension": true,
"status": "ready"
}
2. get_pr_comments
Retrieves comments from an Azure DevOps Pull Request with optional status filtering.
Input:
pr_url(required): Full Azure DevOps PR URL- Format:
https://dev.azure.com/{org}/{project}/_git/{repo}/pullrequest/{id} - Legacy format:
https://{org}.visualstudio.com/{project}/_git/{repo}/pullrequest/{id}
- Format:
status_filter(optional): Filter by comment status- Options:
active,fixed,closed,wontfix,pending,bydesign,unknown,system
- Options:
Output:
{
"pr_url": "https://dev.azure.com/myorg/myproject/_git/myrepo/pullrequest/123",
"status_filter": "active",
"total_comments": 5,
"comments": [
{
"id": 1,
"thread_id": 100,
"author": "John Doe",
"author_email": "john.doe@example.com",
"content": "Please add error handling here",
"status": "active",
"thread_status": "active",
"comment_type": "text",
"published_date": "2025-11-20T10:30:00Z",
"last_updated_date": "2025-11-20T10:30:00Z"
}
]
}
Example PR URL Formats
https://dev.azure.com/myorg/myproject/_git/myrepo/pullrequest/123
https://myorg.visualstudio.com/myproject/_git/myrepo/pullrequest/456
MCP Client Configuration
VS Code with GitHub Copilot (User-Level Global Configuration)
For global access across all workspaces, add the server to your user-level MCP configuration:
File Location: ~/.config/Code/User/mcp.json (Linux/macOS) or %APPDATA%\Code\User\mcp.json (Windows)
Add this entry to the servers object:
{
"servers": {
"simple-azure-devops-mcp": {
"command": "node",
"type": "stdio",
"args": ["/absolute/path/to/azure-mcp/dist/index.js"]
}
}
}
Important: Replace /absolute/path/to/azure-mcp/dist/index.js with the actual path to your built project.
Setup Steps:
- Build the project:
pnpm build - Edit
~/.config/Code/User/mcp.jsonand add the server configuration - Reload VS Code window (Cmd/Ctrl + Shift + P → "Developer: Reload Window")
- Open GitHub Copilot Chat
- The MCP tools will be available in all workspaces
VS Code with GitHub Copilot (Workspace Level)
Create or edit .vscode/settings.json to automatically enable MCP server for the current workspace:
{
"github.copilot.chat.mcp.enabled": true,
"github.copilot.chat.mcp.servers": {
"simple-azure-devops-mcp": {
"command": "node",
"args": ["${workspaceFolder}/dist/index.js"]
}
}
}
Setup Steps:
- Open this workspace in VS Code
- Ensure the project is built:
pnpm build - Reload VS Code window (Cmd/Ctrl + Shift + P → "Developer: Reload Window")
- Open GitHub Copilot Chat
- The MCP tools will be automatically available
Note: Make sure you have GitHub Copilot enabled and MCP support is available in your VS Code version.
VS Code (Cline Extension)
- Install the Cline extension from the VS Code marketplace
- Open Cline Settings (click the gear icon in Cline panel)
- Add MCP Server in the MCP Servers section:
{
"mcpServers": {
"simple-azure-devops-mcp": {
"command": "node",
"args": ["/absolute/path/to/azure-mcp/dist/index.js"]
}
}
}
Important: Use the absolute path to your project. For example:
- Linux/macOS:
/home/username/Code/azure-mcp/dist/index.js - Windows:
C:\\Users\\username\\Code\\azure-mcp\\dist\\index.js
- Restart Cline or reload VS Code
- The tools should now appear in Cline's available tools
VS Code Settings Location
You can also manually edit the Cline configuration file:
- Linux:
~/.config/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json - macOS:
~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json - Windows:
%APPDATA%\Code\User\globalStorage\saoudrizwan.claude-dev\settings\cline_mcp_settings.json
Claude Desktop
Add this server to your Claude Desktop configuration:
{
"mcpServers": {
"simple-azure-devops-mcp": {
"command": "node",
"args": ["/absolute/path/to/azure-mcp/dist/index.js"]
}
}
}
Configuration file location:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
Development
This project uses pnpm as its official package manager. Please use pnpm for all package management operations to ensure consistency and proper dependency resolution.
Run tests
pnpm test
Run tests with coverage
pnpm test:coverage
Linting and Formatting
This project uses ESLint for linting and Prettier for code formatting:
# Run linter
pnpm lint
# Fix linting issues automatically
pnpm lint:fix
# Format code with Prettier
pnpm format
# Check formatting without making changes
pnpm format:check
Git Hooks
Husky is configured to run lint-staged on pre-commit, which automatically:
- Runs ESLint with auto-fix on staged TypeScript files
- Formats code with Prettier
This ensures code quality and consistency before commits.
Watch mode for development
pnpm dev
Project Structure
azure-mcp/
├── src/
│ ├── index.ts # Main MCP server
│ ├── azure-cli.ts # Azure CLI utilities
│ ├── pr-comments.ts # PR comment retrieval logic
│ ├── azure-cli.test.ts # Tests for Azure CLI utilities
│ └── pr-comments.test.ts # Tests for PR comment parsing
├── dist/ # Compiled JavaScript (generated)
├── package.json # Package configuration
├── tsconfig.json # TypeScript configuration
├── vitest.config.ts # Vitest test configuration
└── README.md # This file
Comment Status Types
- active: Comment thread is active and requires attention
- fixed: Issue has been addressed
- closed: Thread is closed
- wontfix: Issue won't be addressed
- pending: Waiting for action
- bydesign: Behavior is intentional
- unknown: Status is not determined
- system: System-generated comment
Troubleshooting
Azure CLI not found
Error: Azure CLI is not installed or not available in PATH
Solution: Install Azure CLI and ensure it's in your PATH.
Not authenticated
Error: Not authenticated. Please run: az login
Solution: Run az login and complete the authentication flow.
Azure DevOps extension missing
The server will automatically attempt to install the azure-devops extension. If this fails:
az extension add --name azure-devops
Invalid PR URL format
Error: Invalid Azure DevOps PR URL format
Solution: Ensure your PR URL matches one of these formats:
https://dev.azure.com/{org}/{project}/_git/{repo}/pullrequest/{id}https://{org}.visualstudio.com/{project}/_git/{repo}/pullrequest/{id}
Contributing
- Fork the repository
- Create a feature branch
- Make your changes
- Run tests:
pnpm test - Submit a pull request
License
MIT
Resources
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。