Most enterprises aren’t short on AI ambition. What they’re missing is a secure, reliable way to move from experimentation to execution, especially as generative models and agentic workflows start to touch production systems. CrowdStrike’s latest move with AWS addresses this gap head-on.
Through an expanded collaboration announced at
AWS Summit NYC on June 16th, CrowdStrike made two key offerings - falcon-mcp and AI Red Team Services - available in the new AI Agents and Tools category of AWS Marketplace. These additions give AWS customers a way to both embed AI into their security operations and validate the AI systems themselves.
Secure AI Workflows with falcon-mcp
falcon-mcp acts as a middleware layer between AI agents and the CrowdStrike Falcon platform, making it easier to connect models to live telemetry, threat intelligence, and behavioral data. Delivered as a managed control plane (MCP) server, it supports standardized, open protocols for integrating AI-powered workflows securely. Available via Amazon Bedrock AgentCore, falcon-mcp is designed to accelerate deployment and reduce complexity by offering direct access to Falcon data streams.
"falcon-mcp is how CrowdStrike will help organizations move fast and build securely in the age of agentic AI,” said
Daniel Bernard, chief business officer at CrowdStrike. "It gives customers a simple, secure way to connect AI agents to the same real-time telemetry that protects some of the most critical infrastructure in the world.”
"What makes our approach different is that we’re not layering AI on top – we’re extending the core of the Falcon platform to support secure automation by design. That gives customers and partners a trusted foundation to build AI-powered workflows, with confidence in the security behind every decision.”
Pressure-Test GenAI with AI Red Team Services
Alongside workflow enablement, CrowdStrike is also focusing on validation. Its AI Red Team Services allow organizations to test LLMs and GenAI applications for security weaknesses, including misconfigurations, prompt injection paths, and over-permissive access. These services align with frameworks like the OWASP Top 10 for LLMs and offer tailored guidance to strengthen both models and infrastructure.
"The biggest risks we see today come down to speed outpacing security,” Bernard added. “Many early GenAI deployments are giving AI systems too much access, without enough oversight. That opens the door to unintended actions, data exposure, and abuse. Identity protection is also among the top agentic risks, as agent espionage and misuse are becoming effective and successful.”
"Our AI Red Team services are helping customers identify those gaps early and put guardrails in place before they become incidents. The insights from that work are shaping best practices across the AWS ecosystem – so organizations can embrace AI confidently and with greater speed. Successful AI adoption starts with foundational security. Cybersecurity is mission-critical in an AI-accelerated world.”
Bringing It All Together with AWS
Together, these tools make it easier for enterprises to develop, test, and operate AI securely on AWS. CrowdStrike already integrates across AWS services like SageMaker and Bedrock, and this latest move deepens that alignment. As organizations begin embedding agentic AI into customer service, security operations, and cloud workflows, the focus is shifting from “how fast can we build” to “how safely can we scale.” CrowdStrike is betting that the answer starts with visibility, integration, and validation - available now through a familiar AWS Marketplace experience.
