This MCP server enables AI assistants to interact with Azure DevOps data through a set of specialized tools. It exposes workers for Work Items, Git Repositories, Pull Requests, and Pipelines, allowing you to fetch, create, update, and search items, query WIQL, browse repository contents, inspect branches, manage PRs, and inspect pipeline runs. Use it with any MCP client that supports the protocol (e.g., Claude Code/Desktop) to integrate Azure DevOps data into conversations or automated workflows. Start by running the server (via Docker in production or with the .NET CLI for development), then query endpoints like get_work_item, query_work_items, get_repositories, get_pull_requests, and get_pipelines to retrieve structured data and perform operations through natural language commands or structured requests.

Prerequisites:

• Docker and Docker Compose installed on your machine
• Access to an Azure DevOps organization and a Personal Access Token (PAT)
• Optional: .NET SDK for local development (not required if using Docker)

Installation steps:

1. Clone the repository and navigate to it: git clone https://github.com/viamus/mcp-azure-devops.git cd mcp-azure-devops

2. Prepare environment variables for Azure DevOps access. Create a .env file or use the provided .env.example to set: AZURE_DEVOPS_ORG_URL=https://dev.azure.com/your-organization AZURE_DEVOPS_PAT=your-personal-access-token AZURE_DEVOPS_DEFAULT_PROJECT=your-project-name

3. Start the server (Docker, recommended): docker compose up -d This will run the MCP server and expose the health endpoint at http://localhost:8080/health (internal default).

4. (Optional) Run locally with .NET CLI (development): dotnet run --project src/Viamus.Azure.Devops.Mcp.Server Server will be available at http://localhost:5000

5. Verify the server is healthy: curl http://localhost:8080/health

   or if using .NET CLI: curl http://localhost:5000/health

• The Docker setup uses a .env-based configuration for credentials. Ensure you set the AZURE_DEVOPS_* environment variables before starting.
• If you are behind a reverse proxy or need HTTPS, configure TLS termination at the proxy rather than inside the MCP container.
• The PAT provided to Azure DevOps should have at least Read/Write permissions for Work Items, Code, and Build as described in the PAT guidance.
• For Windows automation, there is an optional PowerShell installer script; run it with proper admin privileges and provide your Azure DevOps details as shown in the README.
• The MCP server exposes a set of tools for different Azure DevOps capabilities. Use the tools in your MCP client by invoking named actions (e.g., get_work_item, create_work_item, get_repositories, get_pull_requests, query_builds).
• If you plan to deploy to production, consider configuring a reverse proxy, enabling HTTPS, and securing API access with an API key if supported by your MQTT client (refer to your client’s docs).