TechPrufer
Get started
Microsoft
Microsoft logoMicrosoft · Feature trail

Azure DevOps MCP Server

10 posts · Dec 2022 – Aug 2026

What this trail covers

This thread tracks the evolution of the Azure DevOps MCP Server, which provides AI assistants with secure, contextual access to Azure DevOps projects for planning, building, and shipping software. The server runs locally, accessing private data without it leaving the system, and supports work items, wiki, repos, and search. This update announces the general availability of the local MCP Server, signifying increased stability and a more deliberate update cadence for the open-source project. Previous updates focused on enabling AI assistants to interact with Azure DevOps projects for tasks like code generation, commit message creation, and pull request summarization. This post extends the capability by highlighting how GitHub Copilot, an AI assistant, can be leveraged by Azure DevOps users, integrating with tools like Visual Studio and VS Code, and providing AI-driven coding assistance and automation within Azure Repos, including commit messages, pull request descriptions, and Azure DevOps Wiki content generation.

All Microsoft trails
  1. Azure DevOps Remote MCP Server is generally availableAnnounces the general availability of the Azure DevOps Remote MCP Server, a hosted endpoint that provides AI assistants secure, contextual access to Azure DevOps projects. It details the setup process using `mcp.json`, authentication via Microsoft Entra (requiring organizations to be backed by a Microsoft Entra tenant), and lists supported clients like Visual Studio Code with GitHub Copilot, Microsoft Foundry, Microsoft Copilot Studio, Visual Studio, GitHub Copilot CLI, and the GitHub Copilot app. It also mentions that clients requiring dynamic OAuth client registration or client ID Metadata documents in Microsoft Entra are still being worked on, and for those, the local Azure DevOps MCP Server can be used.
  2. Azure DevOps MCP Server April UpdateThis post introduces WIQL query support for work items via a new `wit_query_by_wiql` tool, with limited access for Insiders in the remote MCP. Remote MCP server enhancements include annotations for tool safety, addition of `repo_get_file_content`, `repo_list_directory`, and `repo_vote_pull_request` tools, and restructuring of wiki tools into `wiki` and `search_wiki`. Local MCP server improvements include Personal Access Token (PAT) support for authentication, experimental elicitation prompts for guided input, and experimental MCP Apps for packaging and executing common workflows.
  3. Remote MCP Server preview in Microsoft FoundryIntroduces the Azure DevOps MCP Server (preview) into Microsoft Foundry, a platform for building and managing AI-powered applications and agents. This integration allows agents to connect to Azure DevOps for workflows that enhance developer and product manager productivity. The post details how to add the Azure DevOps MCP Server tool to an agent within Foundry, including selecting it from the catalog, connecting it to an organization, and configuring specific tools for the agent.
  4. Azure DevOps Remote MCP Server (public preview)Introduces the Remote Azure DevOps MCP Server, a hosted version of the local MCP Server that uses streamable HTTP transport and Microsoft Entra for authentication. This removes the need for local setup and installation, supporting Visual Studio and Visual Studio Code with GitHub Copilot. It also notes that standalone organizations (MSAs) are not supported and outlines future client and service support.
  5. Azure Boards integration with GitHub Copilot includes custom agent supportIntroduces support for custom agents within the GitHub Copilot Coding Agent for Azure Boards. Custom agents are defined using Markdown profiles and can be set at the repository or organization level. These agents are automatically available in Azure DevOps for generating code changes and creating pull requests from work items.
  7. Azure Boards integration with GitHub CopilotThis post announces the general availability of the Azure Boards integration with GitHub Copilot, allowing users to create pull requests directly from Azure Boards work items. It introduces the workflow of sending work item details to the Copilot coding agent, which generates a branch and a draft pull request linked back to the work item. New features include branch selection for Copilot and the display of Copilot status on Kanban cards. Future enhancements will include support for custom agents/models and REST API integration. Requirements include GitHub repository hosting, enabled Azure Boards-GitHub integration, and enabled Copilot coding agent.
  8. Azure DevOps and GitHub Repositories — Next Steps in the Path to Agentic AIThis post announces the general availability of the Azure DevOps MCP Server, enabling custom agents defined at the GitHub repository or organization level. It also details the integration of Azure Boards with the GitHub Coding Agent for automated coding tasks and spec-driven development, and the expansion of GitHub Copilot's agentic capabilities to Azure Repos within Visual Studio and VS Code, including AI-assisted code reviews and profiling tools. The post also mentions the availability of GitHub Enterprise Importer for migrating repositories from Azure DevOps to GitHub.
  9. Azure DevOps local MCP Server is generally availableThis post announces the general availability (GA) of the local Azure DevOps MCP Server, moving it out of preview. Key improvements since preview include enhanced login and authorization, refined tooling, and the introduction of 'domains' to scope tools and manage client limits. The GA status implies a commitment to stability and a more deliberate update process for the open-source project.
  10. Azure Boards integration with GitHub Copilot (Private Preview)This post announces the private preview of an integration between Azure Boards work items and the GitHub Copilot coding agent. It allows users to create work items in Azure Boards and send them to GitHub Copilot for tasks like fixing bugs, implementing features, improving test coverage, updating documentation, and addressing technical debt. The integration involves clicking a 'Create a pull request with GitHub Copilot' button on a work item, which passes relevant content from the work item (description, repro steps, last 50 comments) to the Copilot agent. Copilot then generates a branch and a draft pull request linked back to the work item. Requirements include GitHub repos, enabled Azure Boards-GitHub integration, enabled Copilot coding agent in GitHub, and organization approval for the private preview. It also mentions upcoming refinements for backlog levels and GitHub Enterprise Cloud with Data Residency.
  11. From Manual Testing to AI-Generated Automation: Our Azure DevOps MCP + Playwright Success StoryThis post details the integration of the Azure DevOps MCP server with GitHub Copilot and Playwright to automate the generation and execution of end-to-end tests. It outlines a step-by-step process for converting manual test cases into automated Playwright scripts, emphasizing the importance of prompt engineering and providing specific prompt examples. The post also discusses challenges related to handling non-textual test steps and the quality of context provided to the AI. It highlights the benefits of this AI-driven workflow, including faster test creation, broader coverage, seamless CI/CD integration, and on-demand execution.
  12. Azure DevOps MCP Server, Public PreviewThis post announces the public preview availability of the local Azure DevOps MCP Server. It details the server's function as a Model Context Provider that injects real-time Azure DevOps context (work items, pull requests, test plans, etc.) into AI assistants like GitHub Copilot. It emphasizes local execution for data privacy and provides a high-level overview of the setup process, including CLI login and configuration file setup, directing users to the GitHub repository for full instructions and feedback.
  13. GitHub Copilot for Azure DevOps usersThis post announces that GitHub Copilot for Business is accessible to all Azure DevOps customers and highlights its integration with Visual Studio and VS Code. It details how Copilot can be used with Azure Repos for code suggestions, autocompletions, inline comments, automating commit messages, and generating pull request descriptions. It also mentions using Copilot for generating and editing content for Azure DevOps Wiki pages and summarizing changes for commit messages. The post provides links to resources for getting started with GitHub Copilot and using it with Visual Studio and VS Code.
  14. October patches for Azure DevOps ServerThis post details the October patches for Azure DevOps Server, including fixes for bugs in pipeline execution, identity display, and build cleanup jobs. For Azure DevOps Server 2022.0.1 Patch 4, a bug causing pipelines to get stuck was fixed by upgrading the pipeline execution model. Additionally, the build cleanup job was modified to ignore task failures and clean up as many artifacts as possible. For Azure DevOps Server 2020.1.2 Patch 9 and Azure DevOps Server 2020.0.2 Patch 5, a bug where the 'Analysis Owner' identity showed as Inactive on patch upgrade machines was resolved.
  15. Updated: Now available: Azure DevOps Server 2022 RTWThis post details the upgrade path for Azure DevOps Server 2022 RTW, including specific instructions for upgrading Elasticsearch versions (2.x, 5.x, and 6.x to 7.x) and the requirement to select 'Install New Search' for certain upgrade scenarios. It also provides a version mapping between Azure DevOps Server/Team Foundation Server and Elasticsearch.