home / mcp / falcon mcp server
Provides programmatic access to CrowdStrike Falcon capabilities including detections, incidents, intel, and identity protection for agent workflows.
Configuration
View docs{
"mcpServers": {
"itsjahaziel-falcon-mcp-ai": {
"command": "falcon-mcp",
"args": [],
"env": {
"FALCON_BASE_URL": "https://api.crowdstrike.com",
"FALCON_MCP_HOST": "127.0.0.1",
"FALCON_MCP_PORT": "8000",
"FALCON_CLIENT_ID": "your-client-id",
"FALCON_MCP_MODULES": "detections,incidents,intel,spotlight,idp",
"FALCON_CLIENT_SECRET": "your-client-secret",
"FALCON_MCP_TRANSPORT": "stdio"
}
}
}
}You deploy falcon-mcp to connect AI agents with CrowdStrike Falcon, providing programmatic access to security capabilities and enabling automated investigations, detections analysis, and incident workflows within your agentic processes.
You run the Falcon MCP Server locally or in a container and connect your MCP client to it. Start with the default stdio transport to interact directly from your environment, or use SSE/streamable-http transports to enable remote access. You can enable specific modules to tailor the server to your security workflow, such as detections, incidents, intel, and identity protection. Use the client to perform searches, fetch details, and drive automated security tasks from your agent.
Prerequisites: ensure you have Python 3.11 or higher and a way to install Python packages.
Install using the provided package manager for your environment.
Option 1: Install with uv (recommended for development and running the server directly)
uv tool install falcon-mcpOption 2: Install with pip (a straightforward Python packaging approach)
pip install falcon-mcpNote: If the falcon-mcp command is not found, update your PATH to include the installation location.
Run the server with the default stdio transport to begin interaction from your shell or IDE.
Configuration overview: you control which modules are enabled at startup. By default, all available modules are enabled if you do not specify modules. You can override this with a comma-separated list of modules via the command line or environment variables.
Module configuration notes: you require CrowdStrike API credentials to access the Falcon platform. You typically provide a client ID, client secret, and the base API URL. Optional settings let you choose transports, host/port for HTTP-based transports, and debug mode.
Security and credentials: keep your CrowdStrike API credentials secure. Use environment variables or a .env file to pass sensitive information to the server, and avoid embedding secrets in code or logs.
Container usage: you can run Falcon MCP Server in a pre-built container image or build locally for development. The container supports transports such as stdio, SSE, and streamable-http, and you can mount a .env file for configuration.
Editor/Assistant integration: configure your MCP client to connect to falcon-mcp using supported tooling. You can use your editor's MCP integration to run the server and interact with its endpoints directly from your development environment.
Check connectivity to the Falcon API to verify credentials and reachability.
List modules currently enabled on the Falcon MCP Server.
List all available modules that can be enabled.
Search for detections to understand malicious activity across the environment.
Retrieve detailed information for specific detections by ID.
Search for CrowdStrike Discover applications within the environment.
Find unmanaged assets discovered by managed systems.
Search for hosts/devices within the CrowdStrike environment.
Get detailed information for specified hosts.
Investigate an entity in Identity Protection with timeline and risk assessment.
Show CrowdScore and security posture metrics.
Find and analyze security incidents.
Get comprehensive details for a specific incident.
Search for behaviors to understand suspicious activity.
Get details for specific security behaviors.
Research threat actors tracked by intelligence.
Search indicators and IOCs from threat intel.
Access intelligence reports and publications.
Search weekly sensor usage data.
Search vulnerabilities in serverless functions.
Search vulnerabilities within the environment for risk management.