CrowdStrike Launches Falcon Guardian for AI Agent Runtime Security

At its annual Fal.Con conference in Las Vegas on September 1, CrowdStrike announced the launch of Falcon Guardian, a dedicated runtime security layer aimed at safeguarding AI agents operating at the endpoint level. This innovation comes as AI agents increasingly perform complex tasks autonomously, such as writing code and querying databases.

Falcon Guardian enhances CrowdStrike's existing kernel-level Falcon sensor by creating a control point specifically for AI agent activities. It offers real-time inventory of AI agents within an environment, behavioral monitoring to detect unusual actions, and a risk assessment layer to identify threats like prompt injection and unauthorized data access before they escalate.

The product generates compliance audit trails, which are particularly valuable for financial services and healthcare organizations facing regulatory scrutiny over automated systems. Falcon Guardian integrates seamlessly into existing systems without requiring additional third-party tools or SDKs.

Covering over 1,800 distinct AI applications across nearly 160 million instances, Falcon Guardian represents a significant advancement in AI security. It builds on CrowdStrike's previous work with Falcon AIDR, which focused on securing interactions at the prompt layer, by extending its capabilities to monitor runtime activities.

In addition to Falcon Guardian, CrowdStrike also unveiled Falcon IQ, an agentic automation platform featuring more than 50 AI agents designed to assist security operations teams.