EndpointOps-MCP

EndpointOps-MCP

MCP server for managing Microsoft Intune and Entra ID via the Microsoft Graph API, offering 34 tools for device, app, compliance, security, identity, and reporting operations.

Category
Visit Server

README

EndpointOps-MCP

A full read/write Model Context Protocol (MCP) server for managing Microsoft Intune and Entra ID (Azure AD) through the Microsoft Graph API. It gives AI assistants (Claude Desktop, VS Code Copilot, Cursor, Copilot Studio, etc.) direct, structured access to device management, app management, compliance, security, identity, and reporting operations across your tenant.

The server exposes 34 MCP tools, each accepting an action parameter so a single tool can drive dozens of related Graph operations without needing one MCP tool per endpoint. In total the tools cover 280+ underlying Microsoft Graph operations.

⚠️ This server can create, update, and delete real objects in your tenant (devices, users, groups, policies, apps, etc.). Review the Permissions section and scope the app registration carefully before pointing it at a production tenant.


Features

Core / Connection

  • Interactive and app-only (client credentials) authentication flows
  • Session auth status, sign-in, and one-step connect helpers
  • Graph connectivity test and tenant info
  • Intune tenant overview (device counts, compliance, OS breakdown)
  • discover_graph_operations — self-describing catalog of every action every tool supports

Intune Device & App Management

  • List, search, filter, and inspect managed devices
  • Remote actions: sync, restart, lock, rename, locate, wipe, retire, reset passcode, lost mode, defender scan, log collection, diagnostics collection
  • BitLocker / FileVault recovery key lookups and encryption reports
  • Mobile app CRUD, assignments (required/available/uninstall), install status, discovered/detected apps
  • App configuration (MAM) and app protection policies

Compliance & Configuration

  • Compliance policies (CRUD, assignment, deployment status)
  • Device configuration profiles and Settings Catalog policies
  • ADMX / administrative template policies
  • Endpoint security policies and templates (AV, Firewall, EDR)
  • Security baselines (templates + deployed profile status)
  • Windows Update rings and feature/quality/driver update profiles

Enrollment & Fleet Operations

  • PowerShell scripts and proactive remediations (Windows + macOS shell scripts)
  • Enrollment restrictions, Apple VPP/DEP tokens, Android Enterprise settings
  • Windows Autopilot devices, deployment profiles, and enrollment status page (ESP) profiles
  • Assignment filters and scope tags
  • Intune RBAC role definitions and role assignments

Windows 365 Cloud PC

  • Cloud PC lifecycle (list, restart, reprovision, resize, restore, troubleshoot)
  • Provisioning policies, gallery images, network connections, audit events

Entra ID (Azure AD)

  • Full user lifecycle: CRUD, onboarding/offboarding, licenses, manager assignment, password/session management, bulk operations
  • Group management: security, Microsoft 365, and dynamic groups; members/owners
  • Entra device inventory management (list/search/enable/disable/delete)
  • Conditional Access policies and named locations
  • Identity Protection: risky users, risk detections, sign-in/audit logs, MFA and authentication methods
  • App registrations and enterprise app inspection/management

Tenant Administration & Reporting

  • Organization/domain info, service health, directory roles, subscriptions, security defaults
  • Compliance, malware, encryption, endpoint analytics, app reliability, and license usage reports
  • Async Graph report export jobs (deviceManagement/reports/exportJobs)

Available Tools

# Tool Area
1 authenticate_mcp_session Auth
2 test_connection Auth
3 get_auth_status Auth
4 start_interactive_sign_in Auth
5 complete_interactive_sign_in Auth
6 complete_interactive_login Auth
7 connect_intune_mcp_server Auth
8 get_intune_overview Core
9 manage_intune_devices Intune Devices
10 manage_device_encryption Encryption
11 manage_intune_apps Apps
12 manage_app_config_mam App Protection (MAM)
13 manage_compliance_policies Compliance
14 manage_configuration_profiles Config Profiles
15 manage_settings_catalog Settings Catalog
16 manage_admx_policies ADMX / GPO
17 manage_endpoint_security Endpoint Security
18 manage_security_baselines Security Baselines
19 manage_windows_update Windows Update
20 manage_intune_scripts Scripts & Remediations
21 manage_intune_enrollment Enrollment
22 manage_autopilot Autopilot
23 manage_filters_tags Filters & Scope Tags
24 manage_intune_rbac Intune RBAC
25 manage_cloud_pc Windows 365 Cloud PC
26 manage_entra_users Entra Users
27 manage_entra_groups Entra Groups
28 manage_entra_devices Entra Devices
29 manage_conditional_access Conditional Access
30 manage_identity_protection Identity Protection
31 manage_app_registrations App Registrations
32 manage_tenant_admin Tenant Administration
33 manage_intune_reports Reports & Analytics
34 discover_graph_operations Catalog / Discovery

Each manage_* tool takes an action parameter (e.g. manage_intune_devices(action="wipe", device_id="...", confirm=True)). Call discover_graph_operations at any time to get the full, up-to-date list of actions per tool directly from the running server.

Destructive actions (wipe, delete, retire, etc.) require an explicit confirm=True argument — calling them without it returns a confirmation_required response instead of executing.


Required Graph API Permissions

Grant these Application permissions to your Entra ID app registration (Azure Portal → App registrations → API permissions → Microsoft Graph → Application permissions), then click Grant admin consent.

Intune

DeviceManagementManagedDevices.ReadWrite.All
DeviceManagementConfiguration.ReadWrite.All
DeviceManagementApps.ReadWrite.All
DeviceManagementServiceConfig.ReadWrite.All
DeviceManagementRBAC.ReadWrite.All
DeviceManagementScripts.ReadWrite.All
DeviceManagementManagedDevices.PrivilegedOperations.All

Entra ID

User.ReadWrite.All
Group.ReadWrite.All
Directory.ReadWrite.All
RoleManagement.ReadWrite.Directory
Policy.ReadWrite.ConditionalAccess
IdentityRiskyUser.ReadWrite.All
IdentityRiskEvent.Read.All
AuditLog.Read.All
UserAuthenticationMethod.ReadWrite.All
Application.Read.All
Organization.Read.All

Tenant Administration & Reporting

ServiceHealth.Read.All
ServiceMessage.Read.All
Reports.Read.All

Windows 365

CloudPC.ReadWrite.All

Notes:

  • DeviceManagementManagedDevices.PrivilegedOperations.All is required for high-impact remote actions such as device log collection requests.
  • Some enrollment-restriction write actions additionally require the calling principal (or the app registration's effective context) to hold the Global Administrator or Intune Service Administrator directory role — this is a Graph service-side business rule, not just an application permission.
  • If you only need read-only/reporting access, use the .Read.All variants instead of .ReadWrite.All where available and drop any write-only scopes you don't need.

Prerequisites

  • Python 3.10+ (3.11 recommended)
  • A Microsoft Entra ID tenant with Global Administrator (or delegated equivalent) access to create an app registration
  • An Entra ID app registration configured for client credentials (app-only) auth, with the permissions above granted and admin-consented

Installation

git clone https://github.com/nelladath/EndpointOps-MCP.git
cd EndpointOps-MCP
python -m venv .venv
.venv\Scripts\Activate.ps1
pip install -r requirements.txt

Configure environment variables

Copy env.example to .env and fill in your app registration values:

TENANT_ID=your-tenant-id
CLIENT_ID=your-app-client-id
CLIENT_SECRET=your-app-client-secret
AUTH_MODE=app

AUTH_MODE supports:

  • app — app-only (client credentials) token, no user interaction (default, recommended for automation)
  • delegated — signed-in user token only (device-code or browser flow)
  • hybrid — app-only Graph calls gated behind an optional first-layer interactive user sign-in

See env.example for the full list of options (interactive login mode, token cache path, Graph endpoint overrides).


VS Code Setup (MCP Client)

  1. Install the MCP extension support in VS Code (built into recent Copilot Chat versions) or your preferred MCP-compatible client.
  2. Build the console entry point into your virtual environment (done automatically by pip install -r requirements.txt + the project's pyproject.toml script entry, or run directly via python -m intune_mcp_server).
  3. Add a server entry to your workspace .vscode/mcp.json (a template is included in this repo):
{
    "servers": {
        "EndpointOps-MCP": {
            "type": "stdio",
            "command": "${workspaceFolder}/.venv/Scripts/endpointops-mcp.exe",
            "args": [],
            "env": {
                "TENANT_ID": "",
                "CLIENT_ID": "",
                "CLIENT_SECRET": ""
            }
        }
    }
}
  • On macOS/Linux use ${workspaceFolder}/.venv/bin/endpointops-mcp instead.
  • You can leave env empty and rely on a .env file in the project root instead — the server loads it automatically via python-dotenv.
  • Alternatively, run the server as a plain module without the console script: "command": "python", "args": ["-m", "intune_mcp_server"].
  1. Reload the VS Code window (or use the MCP: List Servers command) and the 34 tools listed above will become available to Copilot Chat.
  2. Run test_connection (or connect_intune_mcp_server) first to verify authentication before using any other tool.

Claude Desktop / other MCP clients

Add an equivalent entry to your client's MCP server configuration, pointing command at the same executable/module and passing the same environment variables.


HTTP Transport (optional)

In addition to stdio (the default, used by most desktop MCP clients), the server can run over Streamable HTTP for hosted scenarios (e.g. Copilot Studio custom connectors):

uvicorn intune_mcp_server.mcp_app:app --host 0.0.0.0 --port 8000

A minimal REST bridge exposing a handful of tools as plain HTTP endpoints is also available in bridge.py (FastAPI) for scenarios that need simple REST calls instead of the native MCP protocol.


Security Notes

  • Never commit .env or the MSAL token cache (.msal_token_cache.bin) — both are already excluded via .gitignore.
  • Prefer AUTH_MODE=app with a tightly-scoped app registration for automation; reserve delegated/hybrid modes for interactive, human-supervised sessions.
  • Destructive actions (wipe, retire, delete) require confirm=True — treat any AI client wired to this server with the same caution as a human holding Global Administrator credentials.
  • Grant only the permissions your use case needs; the lists above are the maximum superset for full read/write coverage, not a minimum requirement.

License

MIT © Sujin Nelladath

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured
E2B

E2B

Using MCP to run code via e2b.

Official
Featured