Name: Azure DevOps MCP Server
Author: Tiberriver256

LABSCO SUMMARY

A blocked pull request can be traced to the log line that blocked it without opening the portal.

get_pull_request_checks reports which policy or pipeline is holding the PR and hands back the run identifiers, and pipeline_timeline plus get_pipeline_log take those identifiers down to a specific log range — a chain that most Azure DevOps integrations stop short of. Two things worth knowing before you commit to it: Microsoft ships its own server for Azure DevOps Services, and this one earns its place mainly on Azure DevOps Server on-premises; and create_commit rejects a request where the same file path appears twice, so edits to one file have to be merged into a single change entry.

What it is

A community server over the Azure DevOps REST APIs that also targets Azure DevOps Server on-premises, including older collections. Most calls take organizationId and projectId, so one connection can reach across an organization rather than a single project.

What you get

Projects listed and inspected, with get_project_details pulling in the process, work item types and teams
Work items created, updated, linked and searched — WIQL or a saved query id through list_work_items, plus tags, area path, iteration path, parent and assignee on write
Comment history on a work item, and manage_work_item_link to add or remove a typed relation between two of them
Repositories listed and described with statistics and refs, file content read at a given version, and hierarchical trees for one repository or every repository in a project
Branches created from an existing branch, commits created from a set of file changes, and list_commits returning file-level diffs alongside the commit list
Pull requests created with reviewers, linked work items and tags; updated, commented on at a file and line, and their threads moved between statuses
get_pull_request_changes returns the changed files with unified diffs and policy evaluation status, and get_pull_request_checks summarises the checks that are blocking it with the pipeline and run ids to chase
Pipelines listed, triggered with a branch, variables, template parameters and stages to skip, and their runs read back
pipeline_timeline gives the stages and jobs of a run, get_pipeline_log fetches one log by id with a start and end line, and download_pipeline_artifact returns the text content of a file inside a run artifact
Wikis created and listed, pages read, created and updated, and search_wiki, search_code and search_work_items query across a project

Requirements

An Azure DevOps account with access to the organization, and Node.js v16 or higher; it runs as npx -y @tiberriver256/mcp-server-azure-devops. AZURE_DEVOPS_ORG_URL is the one required variable — the full organization or collection URL. AZURE_DEVOPS_AUTH_METHOD picks between pat, azure-identity and azure-cli and defaults to azure-identity; pat reads AZURE_DEVOPS_PAT, and a service principal additionally needs AZURE_TENANT_ID, AZURE_CLIENT_ID and AZURE_CLIENT_SECRET. Azure DevOps Server on-premises supports PAT only. AZURE_DEVOPS_DEFAULT_PROJECT saves naming a project on every call; AZURE_DEVOPS_API_VERSION and LOG_LEVEL are optional.

Setup effort

One command plus a key — npx -y @tiberriver256/mcp-server-azure-devops, then supply credentials

See installation options

USE CASES

01

Finding out why a pull request will not merge

Read the checks and policy evaluations blocking it, then follow the pipeline and run ids they report into the stage timeline and the specific log lines.

get_pull_request_checkspipeline_timelineget_pipeline_log

02

Reviewing a pull request without leaving the conversation

Pull the changed files with their unified diffs, leave a comment on a specific file and line, and resolve the thread when it is addressed.

get_pull_request_changesadd_pull_request_commentupdate_pull_request_thread_status

03

Shipping a small change from branch to open PR

Branch from the target, commit the file changes, and open the pull request with reviewers and the work items it closes attached.

create_branchcreate_commitcreate_pull_request

04

Turning a request into a tracked, linked work item

Create it with type, area path, iteration and assignee, then link it to the parent or related item it belongs with.

create_work_itemmanage_work_item_linkupdate_work_item

05

Getting oriented in an unfamiliar project

See the repositories and the tree across their default branches, then read the specific files, so exploring does not mean cloning first.

list_repositoriesget_all_repositories_treeget_file_content

06

Answering where something is documented or discussed

Search code, wiki pages and work items for the same phrase, then open the wiki page that turns out to hold the answer.

search_codesearch_wikisearch_work_itemsget_wiki_page

07

Rerunning a build with different inputs and reading the result

Trigger the pipeline on a branch with variables and template parameters, then read the run and fetch text out of the artifact it produced.

trigger_pipelineget_pipeline_rundownload_pipeline_artifact

46 tools

Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.

`add_pull_request_comment`

Add a comment to a pull request (repositoryId optional; derived from pullRequestId when omitted)

`create_commit`

Create a commit on an existing branch using file changes.

Provide plain branch names (no "refs/heads/").
⚠️ Each file path may appear only once per commit request—combine all edits to a file into a single change entry.
Prefer multiple commits when you have sparse or unrelated edits; smaller focused commits keep review context clear.

🎯 RECOMMENDED: Use the SEARCH/REPLACE format (much easier, no line counting!).

Option 1: SEARCH/REPLACE format (EASIEST) Simply provide the exact text to find and replace:

{
  "changes": [{
    "path": "src/api/services/function-call.ts",
    "search": "return axios.post(apiUrl, payload, requestConfig);",
    "replace": "return axios.post(apiUrl, payload, requestConfig).then(r => { processResponse(r); return r; });"
  }]
}

The server fetches the file, performs the replacement, and generates the diff automatically. No line counting, no hunk headers, no context lines needed!

Option 2: UNIFIED DIFF format (Advanced) If you prefer full control, provide complete unified diffs:

Each patch MUST have complete hunk headers: @@ -oldStart,oldLines +newStart,newLines @@
CRITICAL: Every @@ marker MUST include line numbers. Do NOT use @@ without line ranges.
Include 3-5 context lines before and after changes.
For deletions: --- a/filepath and +++ /dev/null
For additions: --- /dev/null and +++ b/filepath

Example unified diff:

{
  "changes": [{
    "patch": "diff --git a/file.yaml b/file.yaml\n--- a/file.yaml\n+++ b/file.yaml\n@@ -4,7 +4,7 @@ spec:\n spec:\n   type: ClusterIP\n   ports:\n-    - port: 8080\n+    - port: 9090\n       targetPort: http\n"
  }]
}

BEFORE YOU INSTALL

What to know first

Tools you get

46 tools when we connected. get_me and list_organizations work only on Azure DevOps Services, so an on-premises Azure DevOps Server install has 44. Six tools, including create_wiki and list_commits, are undocumented.

Requires

Node.js 16+
AZURE_DEVOPS_ORG_URL in your environment
A way to sign in: Azure Identity (the default) or an Azure CLI login for Azure DevOps Services, or AZURE_DEVOPS_PAT with AZURE_DEVOPS_AUTH_METHOD=pat — the only method on Azure DevOps Server

Where it runs

All 6 documented setups run the server on your machine.

CONTEXT COST

Heavy Standing cost

Rated by the size of the 46 tool definitions that ride along on every turn, not by how many there are

INSTALLATION

Connect your client

Server setup

TERMINAL

npx -y @tiberriver256/mcp-server-azure-devops

MCP Clients

Claude Desktop — Azure Identity, Cursor — Azure Identity

JSON

{
  "mcpServers": {
    "azureDevOps": {
      "command": "npx",
      "args": [
        "-y",
        "@tiberriver256/mcp-server-azure-devops"
      ],
      "env": {
        "AZURE_DEVOPS_ORG_URL": "https://dev.azure.com/<YOUR_ORGANIZATION>",
        "AZURE_DEVOPS_AUTH_METHOD": "azure-identity",
        "AZURE_DEVOPS_DEFAULT_PROJECT": "<YOUR_AZURE_DEVOPS_DEFAULT_PROJECT>"
      }
    }
  }
}

Claude Desktop — access token, Cursor — access token

JSON

{
  "mcpServers": {
    "azureDevOps": {
      "command": "npx",
      "args": [
        "-y",
        "@tiberriver256/mcp-server-azure-devops"
      ],
      "env": {
        "AZURE_DEVOPS_ORG_URL": "https://dev.azure.com/<YOUR_ORGANIZATION>",
        "AZURE_DEVOPS_AUTH_METHOD": "pat",
        "AZURE_DEVOPS_PAT": "<YOUR_PAT>",
        "AZURE_DEVOPS_DEFAULT_PROJECT": "<YOUR_AZURE_DEVOPS_DEFAULT_PROJECT>"
      }
    }
  }
}

Claude Desktop — on-prem, Cursor — on-prem

JSON

{
  "mcpServers": {
    "azureDevOps": {
      "command": "npx",
      "args": [
        "-y",
        "@tiberriver256/mcp-server-azure-devops"
      ],
      "env": {
        "AZURE_DEVOPS_ORG_URL": "https://server:8080/tfs/DefaultCollection",
        "AZURE_DEVOPS_AUTH_METHOD": "pat",
        "AZURE_DEVOPS_PAT": "<YOUR_PAT>",
        "AZURE_DEVOPS_DEFAULT_PROJECT": "<YOUR_AZURE_DEVOPS_DEFAULT_PROJECT>"
      }
    }
  }
}

Azure DevOps MCP Server

A blocked pull request can be traced to the log line that blocked it without opening the portal.

What it is

What you get

Requirements

Setup effort

Finding out why a pull request will not merge

Reviewing a pull request without leaving the conversation

Shipping a small change from branch to open PR

Turning a request into a tracked, linked work item

Getting oriented in an unfamiliar project

Answering where something is documented or discussed

Rerunning a build with different inputs and reading the result

46 tools

add_pull_request_comment

create_branch

create_commit

create_pull_request

create_wiki

create_wiki_page

create_work_item

download_pipeline_artifact

get_all_repositories_tree

get_file_content

get_me

get_pipeline

get_pipeline_log

get_pipeline_run

get_project

get_project_details

get_pull_request

get_pull_request_changes

get_pull_request_checks

get_pull_request_comments

get_repository

get_repository_details

get_repository_tree

get_wiki_page

get_wikis

get_work_item

get_work_item_comments

list_commits

list_organizations

list_pipeline_runs

list_pipelines

list_projects

list_pull_requests

list_repositories

list_wiki_pages

list_work_items

manage_work_item_link

pipeline_timeline

search_code

search_wiki

search_work_items

trigger_pipeline

update_pull_request

update_pull_request_thread_status

update_wiki_page

update_work_item

What to know first

Tools you get

Requires

Where it runs

Heavy Standing cost

Connect your client

Server setup

MCP Clients

Claude Desktop — Azure Identity, Cursor — Azure Identity

Claude Desktop — access token, Cursor — access token

Claude Desktop — on-prem, Cursor — on-prem

Same category

AlibabaCloud DevOps MCP

Azure DevOps .NET MCP Server

Azure DevOps MCP by decriptor

Gitlab MCP Server

`add_pull_request_comment`

`create_branch`

`create_commit`

`create_pull_request`

`create_wiki`

`create_wiki_page`

`create_work_item`

`download_pipeline_artifact`

`get_all_repositories_tree`

`get_file_content`

`get_me`

`get_pipeline`

`get_pipeline_log`

`get_pipeline_run`

`get_project`

`get_project_details`

`get_pull_request`

`get_pull_request_changes`

`get_pull_request_checks`

`get_pull_request_comments`

`get_repository`

`get_repository_details`

`get_repository_tree`

`get_wiki_page`

`get_wikis`

`get_work_item`

`get_work_item_comments`

`list_commits`

`list_organizations`

`list_pipeline_runs`

`list_pipelines`

`list_projects`

`list_pull_requests`

`list_repositories`

`list_wiki_pages`

`list_work_items`

`manage_work_item_link`

`pipeline_timeline`

`search_code`

`search_wiki`

`search_work_items`

`trigger_pipeline`

`update_pull_request`

`update_pull_request_thread_status`

`update_wiki_page`

`update_work_item`