Microsoft Sentinel MCP Server

Name: Microsoft Sentinel MCP Server
Author: dstreefkerk

@dstreefkerk

About Microsoft Sentinel MCP Server

MCP server for Microsoft Sentinel. Enables access to Sentinel logs, incidents, analytics, and Entra ID data via a modular, queryable interface. Strictly non-production. Designed for use with Claude and other LLMs.

#azure #mcp #threat-intelligence #log-analytics #microsoft-sentinel #llm-integration #mcp-server

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "ms-sentinel-mcp-server": {
      "command": "uv",
      "args": [
        "venv"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Microsoft Sentinel MCP Server?

A Model Context Protocol (MCP) server that provides read‑only access to a Microsoft Sentinel instance, including advanced KQL querying, incident viewing, and resource exploration for Azure Sentinel environments. It is designed for observation‑only security operations and analysis in test environments only.

How to use Microsoft Sentinel MCP Server?

Authenticate with Azure CLI (az login), clone the repository, run the PowerShell installation script (.\install.ps1) to create a virtual environment, install dependencies, and generate a client configuration file. Then paste the configuration into an MCP‑compatible client (e.g., Claude Desktop, Cursor). Alternatively, set environment variables for service principal authentication and run manually with mcp run wrapper.py.

Key features of Microsoft Sentinel MCP Server

Execute and validate KQL queries, including mock data testing
List and view Log Analytics workspace tables and schemas
List, view, and analyze security incidents
Access analytics rules, rule templates, and hunting queries
List data connectors, watchlists, and threat intelligence data
View Entra ID users, groups, and RBAC role assignments

Use cases of Microsoft Sentinel MCP Server

Security analysts querying Sentinel incidents via an LLM interface
Testing KQL queries with synthetic data before production use
Exploring Sentinel workspace metadata and table schemas
Auditing analytics rule coverage by MITRE tactics and techniques
Retrieving user and group details from Microsoft Entra ID for access reviews

FAQ from Microsoft Sentinel MCP Server

Is this server safe to connect to a production Sentinel instance?

No. This server is intended only for test environments. Connecting it to a production Sentinel instance or production Microsoft Entra ID may expose sensitive user and directory data to LLM operators or public LLMs. Use only with non‑production/test tenants or a private LLM.

What authentication methods does Microsoft Sentinel MCP Server support?

It supports any method compatible with the Azure Python SDK’s DefaultAzureCredential, including Azure CLI authentication and service principal authentication via environment variables (AZURE_TENANT_ID, AZURE_CLIENT_ID, AZURE_CLIENT_SECRET). The service principal should be assigned roles like Log Analytics Reader and Microsoft Sentinel Reader.

What operations can I perform with Microsoft Sentinel MCP Server?

All operations are read‑only. You can run KQL queries, list and view incidents, analytics rules, rule templates, hunting queries, data connectors, watchlists, metadata, source controls, ML analytics settings, Entra ID users/groups, and threat intelligence (domain WHOIS, IP geolocation). No write or modify actions are available.

What are the prerequisites for using Microsoft Sentinel MCP Server?

You need a Microsoft Sentinel workspace in Azure (test tenant only), Python installed, and either the Azure CLI authenticated or service principal credentials configured. The server also requires a Log Analytics workspace with the Sentinel resource provider registered.

Does Microsoft Sentinel MCP Server support streaming or SSE transports?

Yes, the server can be run in SSE mode for use with IDEs by executing python wrapper.py --sse. It also supports the standard MCP stdio transport and can be run in development mode with hot reload (mcp dev wrapper.py).