Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Azure DevOps MCP Server list all active bugs assigned to me"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP (Model Context Protocol) server for managing Azure DevOps Work Items including Epics, Features, User Stories, Tasks, Bugs, and more.
Read Work Items: Get individual work items or list by type with filters
Create Work Items: Create Epics, Features, User Stories, Tasks, Bugs, etc.
Update Work Items: Modify any field including state, assignment, description
Delete Work Items: Move to recycle bin or permanently delete
Query Work Items: Execute custom WIQL queries
Link Work Items: Create parent-child and other relationships
Get Metadata: List work item types, iterations, and area paths
Related MCP server: Azure DevOps MCP Server
Node.js 18 or higher
Azure DevOps account with a Personal Access Token (PAT)
PAT needs Work Items (Read & Write) scope
npm install
npm run buildSet the following environment variables:
export AZURE_DEVOPS_ORG="your-organization"
export AZURE_DEVOPS_PROJECT="your-project"
export AZURE_DEVOPS_PAT="your-personal-access-token"Go to Azure DevOps > User Settings > Personal Access Tokens
Create new token with these scopes:
Work Items: Read & Write
Copy the token (you won't see it again!)
Add to your Claude Desktop configuration (claude_desktop_config.json):
{
"mcpServers": {
"azure-devops": {
"command": "node",
"args": ["path/to/azure-devops-mcp/dist/index.js"],
"env": {
"AZURE_DEVOPS_ORG": "your-organization",
"AZURE_DEVOPS_PROJECT": "your-project",
"AZURE_DEVOPS_PAT": "your-pat"
}
}
}
}Get a single work item by ID.
{
"id": 123,
"expand": "Relations"
}List work items by type with optional filters.
{
"type": "User Story",
"state": "Active",
"assignedTo": "user@example.com",
"top": 10
}Create a new work item.
{
"type": "Task",
"title": "Implement login feature",
"description": "<p>Details here</p>",
"assignedTo": "developer@example.com",
"priority": 2,
"parentId": 100
}Update an existing work item.
{
"id": 123,
"state": "Resolved",
"comment": "Fixed the issue"
}Delete a work item.
{
"id": 123,
"permanent": false
}Execute a WIQL query.
{
"query": "SELECT [System.Id] FROM WorkItems WHERE [System.State] = 'Active' AND [System.AssignedTo] = @Me"
}Get all children of a work item.
{
"parentId": 100
}Link two work items.
{
"sourceId": 100,
"targetId": 101,
"linkType": "System.LinkTypes.Hierarchy-Forward",
"comment": "Added as child"
}List available work item types.
List project iterations/sprints.
List project area paths.
System.LinkTypes.Hierarchy-Forward - Parent to Child
System.LinkTypes.Hierarchy-Reverse - Child to Parent
System.LinkTypes.Related - Related work items
System.LinkTypes.Dependency-Forward - Successor
System.LinkTypes.Dependency-Reverse - Predecessor
System.Title - Title
System.Description - Description (HTML)
System.State - State (New, Active, Resolved, Closed)
System.AssignedTo - Assigned user
System.AreaPath - Area path
System.IterationPath - Iteration/Sprint
System.Tags - Tags (semicolon separated)
Microsoft.VSTS.Common.Priority - Priority (1-4)
Microsoft.VSTS.Scheduling.StoryPoints - Story points
Microsoft.VSTS.Scheduling.OriginalEstimate - Original estimate
Microsoft.VSTS.Scheduling.RemainingWork - Remaining work
Microsoft.VSTS.Common.AcceptanceCriteria - Acceptance criteria
MIT
List area paths
Inputs: no named parameters.
Get child work items of a parent
Inputs: parentId (required).
Get current Azure DevOps organization and project configuration
Inputs: no named parameters.
Get the current active sprint/iteration for a team. Returns sprint name, dates, and path.
Inputs: team.
Get the current authenticated user identity (useful for knowing your user ID for voting on PRs, assigning work items, etc.)
Inputs: no named parameters.
Get the permission level and group memberships of the currently authenticated user. Returns whether the user is an org admin, project admin, contributor, or reader. Useful to understand what operations are allowed before attempting them.
Inputs: no named parameters.
List iterations/sprints
Inputs: no named parameters.
Get all work items for a specific iteration/sprint path. Faster than query_work_items for sprint planning.
Inputs: iterationPath (required), type, state, top.
Get your current role in this MCP server. Shows detected role, source (Azure DevOps groups or env var), and what level of access you have.
Inputs: no named parameters.
Get current project details
Inputs: no named parameters.
Get details of a specific repository
Inputs: repositoryId (required).
Get the current sprint/iteration board with all work items assigned to it
Inputs: team.
Get team member capacity for a sprint (hours per day, days off)
Inputs: iterationId (required), team.
Get all work items in a specific sprint iteration.
Inputs: iterationId (required), team.
List all members of a specific team.
Inputs: teamId (required).
List all teams in the current project. Use team names with get_sprint_board.
Inputs: no named parameters.
Get profile information for an Azure DevOps user by their user ID or "me" for the authenticated user.
Inputs: userId (required).
Get details of a specific wiki
Inputs: wikiIdentifier (required).
Get content of a wiki page by path
Inputs: wikiIdentifier (required), path (required), recursionLevel.
Get the revision/edit history of a wiki page (who changed it and when)
Inputs: wikiIdentifier (required), path (required), top.
Get the raw markdown/text content of a wiki page (without metadata).
Inputs: wikiIdentifier (required), path (required).
Get a single work item by ID with all fields and relations
Inputs: id (required), expand.
List all attachments on a work item (files, images, documents).
Inputs: workItemId (required).
Get all comments/discussion history for a work item
Inputs: workItemId (required), top, order.
Get the parent work item of a given work item. Returns null if the work item has no parent.
Inputs: workItemId (required).
Fetch multiple work items by their IDs in a single request. More efficient than calling get_work_item repeatedly.
Inputs: ids (required), fields, expand.
Get detailed information about a work item type including its fields, states, and icon.
Inputs: workItemType (required).
List available work item types
Inputs: no named parameters.
List all workflow states for a given work item type (e.g. Bug, Task, User Story). Useful to know valid state transitions.
Inputs: workItemType (required).
Get the full change history of a work item (who changed what and when)
Inputs: workItemId (required), top.
List team dashboards. Returns dashboard names and IDs.
Inputs: team.
List all security groups defined in the current project.
Inputs: no named parameters.
List all Git repositories in the project
Inputs: no named parameters.
List saved work item queries in the project
Inputs: top.
List all work item tags in the project
Inputs: no named parameters.
List all pages in a wiki (table of contents)
Inputs: wikiIdentifier (required), path, recursionLevel.
List all wikis in the project
Inputs: no named parameters.
List all available work item field names and their reference names (useful for custom queries)
Inputs: no named parameters.
List all available work item relation/link types (use these values with link_work_items)
Inputs: no named parameters.
List work items by type (Epic, Feature, User Story, Task, Bug) with filters
Inputs: type (required), state, assignedTo, areaPath, +2 more.
Execute a WIQL query. Example: SELECT [System.Id] FROM WorkItems WHERE [System.WorkItemType] = 'Bug' AND [System.State] = 'Active'
Inputs: query (required), top.
Execute a saved work item query and return the matching work items
Inputs: queryId (required).
Search for code across all repositories in the project. Returns matching files with line context.
Inputs: searchText (required), top, repository, branch.
Search work items by text across title, description and comments. Easier than WIQL for natural language queries.
Inputs: searchText (required), top, skip.
Switch to a different project within the current organization. Use list_projects first to see available projects.
Inputs: project (required).
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 45 tools
Most tools have distinct purposes with clear resource-action pairs, such as get_work_item, list_work_items, and query_work_items. However, some overlap exists, like get_sprint_work_items and get_iteration_work_items, which could cause confusion despite descriptive names. Overall, the set is well-differentiated with only minor ambiguities.
Tool names follow a highly consistent verb_noun pattern throughout, using verbs like get, list, search, set, and run. All names use snake_case uniformly, making them predictable and easy to parse. This consistency aids agent selection and reduces cognitive load.
With 45 tools, the count is high and may feel heavy for the Azure DevOps domain, potentially overwhelming agents. While the tools cover many aspects, some could be consolidated or omitted without losing functionality. It's borderline excessive but still manageable given the comprehensive scope.
The tool surface provides extensive coverage for Azure DevOps, including work items, repositories, wikis, sprints, queries, and user management. It supports full CRUD-like operations (e.g., get, list, query, search) across all major resources, with no obvious gaps that would hinder agent workflows in this domain.
Crie épicos, features, histórias e tasks no Azure DevOps a partir de uma conversa.
Laviya project management: read and update tasks, wikis, sprints and test cases.
Manage Loom projects, tasks, bugs, defects, clients and phases from AI assistants
Task manager your agent can fully operate: boards, tasks, sprints, roles, worklogs, day planner.