MCP server/cloud

adtk MCP Server

The most comprehensive Azure DevOps MCP server & CLI

★ 1zach-snell/adtk ↗by zach-snellupdated
1

Add it to Claude Code

claude mcp add -e "AZURE_DEVOPS_ORG=${AZURE_DEVOPS_ORG}" -e "AZURE_DEVOPS_PAT=${AZURE_DEVOPS_PAT}" adtk -- adtk mcp
Required:AZURE_DEVOPS_ORGAZURE_DEVOPS_PAT+ 1 optional
2

Make your agent remember this setup

adtk's config, env vars, and the gotchas you hit — recalled in every future Claude Code, Cursor, and Codex session.

npx conare@latest

Free · one command · indexes the sessions already on disk. Set up in the browser instead →

What it does

  • 14 consolidated MCP tools with 91 actions covering Azure DevOps domains
  • Built-in support for boards, iterations, test plans, and security alerts
  • Git branch detection to auto-detect work item IDs from branch names
  • Work item metrics including cycle time, lead time, and time-in-status
  • Token-optimized responses for AI agents

Tools 1

azure-devops-toolsA collection of 14 consolidated tools covering Azure DevOps domains including projects, work items, repositories, pull requests, pipelines, boards, and test plans.

Environment Variables

AZURE_DEVOPS_ORGrequiredThe Azure DevOps organization name
AZURE_DEVOPS_PATrequiredPersonal Access Token for authentication
ADTK_ENABLE_WRITESGate for enabling mutation operations

Try it

Summarize the current sprint progress for the project.
Review the latest pull requests in the main repository and provide a summary.
Check the health status of the current CI/CD pipelines.
List all active bugs assigned to me in the current iteration.
Calculate the cycle time for work item 42.
Original README from zach-snell/adtk

adtk — Azure DevOps Toolkit

A dual-mode Go CLI & MCP server for Azure DevOps. Single binary, PAT-first auth, 14 MCP tools with 91 actions, 4 MCP prompts.

The most comprehensive Azure DevOps MCP server. A single Go binary — CLI for humans, MCP for AI agents.

Comparison: adtk vs Microsoft azure-devops-mcp

Feature adtk microsoft/azure-devops-mcp
Language Go (single static binary) TypeScript (Node.js)
MCP Tools / Actions 14 tools / 91 actions ~75 individual tools
MCP Prompts 4 built-in prompts None
CLI mode Full CLI with 15 command groups No
Auth PAT (self-service, no admin) Azure AD (requires admin consent)
Startup ~50ms ~2s
Response flattening System.Titletitle Raw API responses
Boards & Iterations Full support No
Test Plans Full support No
Advanced Security Alert listing & details No
Metrics Cycle time, lead time, time-in-status No
Git branch detection Auto-detect work items from branch No
Branch policies & tags List policies, list/create tags No
Variable groups & environments List/get variable groups, environments No
Saved queries Get and run saved queries No
Attachments Upload, download, list No
Write protection ADTK_ENABLE_WRITES gate None
Rate limiting Built-in token bucket None
Binary size ~15 MB npm install (~200+ MB)

Features

  • Dual-mode — Full CLI with table output + MCP server for AI agents
  • 14 consolidated MCP tools with 91 actions covering every Azure DevOps domain
  • 4 MCP prompts — sprint_summary, pr_review_digest, pipeline_health, release_readiness
  • Git branch detection — Auto-detect work item IDs from branch names (e.g., feature/12345-description)
  • Work item metrics — Cycle time, lead time, time-in-status computed from revision history
  • PAT-first auth — Self-service Personal Access Tokens, no Azure AD admin approval
  • Single binary — No Node.js, Python, or Docker required
  • Response flattening — Strips _links and converts System.* fields to readable names
  • Write protection — All mutations gated behind ADTK_ENABLE_WRITES=true
  • Rate limiting — Built-in token bucket respecting Azure DevOps TSTU limits
  • Token-optimized — AI agents get clean, concise payloads (40-60% fewer tokens)

Installation

go install

go install github.com/zach-snell/adtk/cmd/adtk@latest

Build from source

git clone https://github.com/zach-snell/adtk.git
cd adtk
./install.sh

Pre-built binaries

Download from the Releases page.

Quick Start

Authenticate

adtk auth
# Or use environment variables:
export AZURE_DEVOPS_ORG=myorg
export AZURE_DEVOPS_PAT=your-pat-here

CLI Usage

# Projects
adtk projects list
adtk projects get MyProject
adtk projects teams MyProject

# Work items
adtk work-items get 42
adtk work-items list -p MyProject

# Repositories
adtk repos list -p MyProject
adtk repos branches myrepo -p MyProject
adtk repos tree myrepo /src -p MyProject
adtk repos policies myrepo -p MyProject
adtk repos tags myrepo -p MyProject

# Pull requests
adtk pull-requests list myrepo -p MyProject
adtk pull-requests get myrepo 1

# Pipelines
adtk pipelines list -p MyProject
adtk pipelines runs 42 -p MyProject
adtk pipelines var-groups -p MyProject
adtk pipelines var-group 1 -p MyProject
adtk pipelines environments -p MyProject

# Iterations & boards
adtk iterations current -p MyProject
adtk boards list -p MyProject
adtk boards columns Stories -p MyProject

# Wiki
adtk wiki list -p MyProject
adtk wiki get ProjectWiki /Home -p MyProject

# Work item metrics
adtk work-items metrics 42 -p MyProject

# Work item auto-detect from git branch
adtk work-items get          # auto-detects work item ID from branch name

# Search
adtk search code "func main" -p MyProject
adtk search work-items "login bug" -p MyProject
adtk search wiql "SELECT [System.Id] FROM WorkItems WHERE [System.State] = 'Active'"
adtk search query "My Saved Queries/Active Bugs" -p MyProject

# Test plans
adtk test-plans list -p MyProject

# Security alerts
adtk security alerts myrepo -p MyProject

# Attachments
adtk attachments list 42 -p MyProject

# All commands support --json for raw output
adtk projects list --json

MCP Server

stdio m

Frequently Asked Questions

What are the key features of adtk?

14 consolidated MCP tools with 91 actions covering Azure DevOps domains. Built-in support for boards, iterations, test plans, and security alerts. Git branch detection to auto-detect work item IDs from branch names. Work item metrics including cycle time, lead time, and time-in-status. Token-optimized responses for AI agents.

What can I use adtk for?

Automating sprint summaries and pipeline health reports for engineering teams. Streamlining pull request reviews by providing AI-generated digests. Tracking work item metrics and cycle times directly from Azure DevOps data. Managing and querying work items, boards, and test plans via natural language. Searching across codebases and work items using AI-assisted queries.

How do I install adtk?

Install adtk by running: go install github.com/zach-snell/adtk/cmd/adtk@latest

What MCP clients work with adtk?

adtk works with any MCP-compatible client including Claude Desktop, Claude Code, Cursor, and other editors with MCP support.

Conare · memory for coding agents

Turn this server into reusable context

Keep adtk docs, env vars, and workflow notes in Conare so your agent carries them across sessions.

Set up free$npx conare@latest