microsoft/azure-devops-mcp

What it solves

This project provides a way for AI agents to interact with Azure DevOps, allowing them to retrieve information and perform actions within the platform. It solves the problem of AI agents lacking direct, structured access to project management and development tools in Azure DevOps.

How it works

The server implements the Model Context Protocol (MCP), which acts as a thin layer over the Azure DevOps REST APIs. This allows AI agents to use specific tools to perform tasks such as listing projects, managing work items, and updating wiki pages. The project offers two deployment options: a hosted remote server for easier setup and a local server for those requiring a stdio connection.

Who it’s for

Developers and teams using Azure DevOps who want to integrate AI agents (such as GitHub Copilot in Agent mode) into their development workflow to automate project tracking, documentation, and repository management.

Highlights

• Broad Toolset: Supports a wide range of Azure DevOps functions including work items, pipelines, repositories, wikis, and test plans.
• Flexible Deployment: Available as a hosted remote server or a local Node.js installation.
• Domain Filtering: The local server allows users to load only specific tool groups (e.g., work-items or pipelines) to keep the tool list manageable for the agent.
• MCP Standard: Built on the Model Context Protocol to ensure compatibility with supported MCP clients like VS Code, Cursor, and Claude Code.