Azure Devops MCP Server
Azure DevOps work items, git repositories, and pipelines for AI assistants.
Data last scanned today · Reviewed today
Overview
This MCP server bridges Azure DevOps with AI interfaces, providing programmatic access to project management and version control data. It allows users to query work items, manage pull requests, and monitor pipelines directly through their preferred AI-powered tools.
Our verdict
With a security score of 35/100, this server is currently high-risk. The primary concern is the lack of a read-only mode, meaning every installation is granted full write access to your Azure DevOps environment. Given the reliance on API key authentication without OAuth support, this tool is best suited for experienced developers working in isolated environments rather than sensitive enterprise setups.
- Zero direct dependencies minimize the software supply-chain surface area.
- Project analytics are derived from raw data rather than opaque tools.
- Supports multi-organization configurations for diverse project workflows.
- Exposes ActivatedDate and ClosedDate for custom cycle time calculations.
- No read-only mode exists for safer, restricted data access.
- Lacks OAuth support for more secure token-based authentication.
- Maintained by community contributors rather than an official vendor.
Tools
| Tool | Description | Risk |
|---|---|---|
| get_work_item | Gets details of a specific work item by ID (supports optional includeRelations parameter) | low |
| get_work_items | Gets multiple work items by IDs (batch retrieval, supports optional includeRelations parameter) | low |
| get_work_item_relations | Gets the normalized list of relationships (Parent, Child, Related, Predecessor, Successor, Tests, Tested By, Hyperlink, Attachment) associated with a work item | low |
| get_work_item_tree | Recursively gets parent and child work items starting from a root work item, up to a specified depth limit. Includes cycle detection | low |
| query_work_items | Queries work items using WIQL (Work Item Query Language) | low |
| get_work_items_by_state | Filters work items by state (Active, New, Closed, etc.) | low |
| get_work_items_assigned_to | Gets work items assigned to a specific user | low |
| get_child_work_items | Gets child work items of a parent work item | low |
| get_recent_work_items | Gets recently changed work items | low |
| search_work_items | Searches work items by title text | low |
Compatibility
| Client | Local | Docker | Remote | Read-only |
|---|---|---|---|---|
| Claude Desktop | ||||
| Cursor | ||||
| VS Code | ||||
| Windsurf | ||||
| ChatGPT |
Frequently asked questions
›Can I connect this server to multiple Azure DevOps organizations?
Yes, you can configure multiple organizations by defining them in an array within your appsettings.json file and selecting the desired organization by name during tool calls.
›Does the server provide built-in analytics for lead time or throughput?
No, pre-built analytics tools have been removed. Instead, the server exposes ActivatedDate and ClosedDate fields on all work items, allowing the AI to calculate metrics like cycle time and throughput directly.
›What are the runtime requirements for this server?
The server requires the .NET 10 SDK to run, which can be managed via the .NET CLI or by using the provided Docker configuration.
›Is there an automated setup option for Windows users?
Yes, Windows users can use the provided PowerShell installation script, which automates environment setup, repository cloning, and Claude Code CLI registration.
›Does the server support authenticated communication?
The server authenticates with Azure DevOps using a Personal Access Token (PAT). Additionally, an optional API key can be configured to secure the server itself.
Badge
Maintain this server? Add the live badge to your README.