ado-pr-kit 0.2.3

AdoPRKit: The complete toolbox for Azure DevOps Pull Requests (MCP)

AdoPRKit

AdoPRKit is a Model Context Protocol (MCP) server that provides a complete toolbox for managing Azure DevOps Pull Requests. Designed for enterprise engineers who want a pragmatic, "zero thinking" set of tools for their agents.

Features

• Create PR: Create new pull requests with title, description, and linked work items.
• Get PR: Retrieve detailed information about a specific PR.
• List PRs: List PRs created by you (default) or where you are a reviewer.
• Update PR: Update PR details or perform actions (abandon, draft, publish, reactivate).
• Comments: Add and retrieve comments on PR threads.
• Diffs: View file diffs and changed files in a PR.

Roadmap & Known Limitations

Compared to standard tools like the GitHub MCP server, this ADO implementation has some missing features we plan to address:

• Inline Code Comments: Currently, comments are added to the PR thread generally. We need to support threadContext to pin comments to specific files and lines.
  • Implementation Check: Use threadContext with filePath, rightFileStart (line/offset), and rightFileEnd.
• Rich Comments (Mentions & Work Items): The current tool only creates plain text comments. We need to support:
  • Mentions: @<User Name> (no special ID needed, just text).
  • Work Item Links: #<Work Item ID> (automatically linked by ADO).
• File Content Access: Currently, we can only view diffs. We need a tool to read the full content of any file in the repo (not just what changed) to provide better code reviews.
• Work Item Management: Basic linking exists, but full creation and management (CRUD) of Azure Boards Work Items is missing.
• Merge Strategies: Explicit support for merging PRs with specific strategies (squash, rebase) is not yet exposed via a tool.
• Voting on PRs: Add support for casting votes (Approve, Reject, etc.).
  • Values: Approved (10), Approved with suggestions (5), No vote (0), Waiting for author (-5), Rejected (-10).
  • API: PATCH .../reviewers endpoint.
• Pipeline Status: Ability to check build/CI status checks on a PR.
  • Implementation Check: Use GET .../pullRequests/{id}/statuses to retrieve CI/Build status.

Quick Start

Prerequisites

• Python 3.10+
• Azure DevOps Personal Access Token (PAT) with Code (Read & Write) and Pull Request Threads (Read & Write) scopes.

⚡️ Quick Start (No Install)

Run directly using uvx (part of uv):

uvx ado-pr-manager

📦 Installation

If you prefer installing it globally or in a venv:

pip install ado-pr-manager
# or
uv pip install ado-pr-manager

🛠️ Editor Configuration

Cursor

1. Go to Cursor Settings > General > MCP.
2. Click Add new MCP server.
3. Select Command type.
4. Name: ado-pr-manager
5. Command: uvx ado-pr-manager (or python3 -m ado_pr_kit.server if installed)
6. Add your Environment Variables (AZDO_ORG_URL, AZDO_PAT, etc.) in the config.

VS Code

1. Install the MCP Support extension.
2. Edit your connection config (e.g. .vscode/mcp.json):

{
  "mcpServers": {
    "ado-pr-manager": {
      "command": "uvx",
      "args": ["ado-pr-manager"],
      "env": {
        "AZDO_ORG_URL": "...",
        "AZDO_PAT": "..."
      }
    }
  }
}

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "ado-pr-manager": {
      "command": "uvx",
      "args": ["ado-pr-manager"],
      "env": {
        "AZDO_ORG_URL": "...",
        "AZDO_PAT": "..."
      }
    }
  }
}

🌐 Remote Usage (SSE)

To use this server with web-based MCP clients, you can deploy it as a remote SSE (Server-Sent Events) server.

1. Entry Point: ado_pr_kit.server_sse:app
2. Command: uvicorn ado_pr_kit.server_sse:app --host 0.0.0.0 --port 8000
3. Deployment:
   • Railway/Render: Deploy this repo, set the Start Command to the above, and add your Environment Variables.
   • Docker: Use the provided Dockerfile but override the entrypoint to run uvicorn.

Clients can then connect via SSE at https://your-app-url.com/sse.

🚀 Hosting

Since AdoPRKit is Docker-ready, you can deploy it anywhere.

Option 1: Render (Free Tier Available)

1. Click the button below.
2. Connect your GitHub repository.
3. Add Environment Variables (AZDO_ORG_URL, AZDO_PAT).

[!NOTE] Cold Start: Free-tier hosting (Render, Railway) may spin down after inactivity. The first request after idle may take 30-60 seconds.

Option 2: Railway

1. Click the button below.
2. Railway will detect the Dockerfile.
3. Add variables in the "Variables" tab.

Option 3: Smithery

Built specifically for MCP.

• Deploy on Smithery

Option 4: Self-Host (e.g., alongside LiteLLM)

If you have a customized server (like for LiteLLM), you can run this via Docker Compose.

1. Copy the docker-compose.yml from this repo.
2. Run:

   export AZDO_ORG_URL="https://dev.azure.com/myorg"
   export AZDO_PAT="your-pat"
   docker-compose up -d
   

3. Your MCP server will be at http://YOUR_SERVER_IP:8000/sse.

Development

1. Clone & Install:

   git clone https://github.com/om-surushe/ado-pr-manager.git
   cd ado-pr-manager
   python3 -m venv .venv
   source .venv/bin/activate
   pip install -e .[dev]
   

2. Test:

   pytest
   

Contributing

We welcome contributions! Please see CONTRIBUTING.md for details on how to submit pull requests, report issues, and contribute to the project.

License

This project is licensed under the MIT License - see the LICENSE file for details.