AI Security Hub
Understand the threats facing AI environments and how to defend against them.
AI Security Essentials
Research and hands-on learning for securing AI.
White Paper
Securing AI Where It Executes: The Endpoint is the New Control Point for AI Agent Security
AI agents are no longer passive tools — they’re autonomous systems that execute commands, modify files, and access sensitive data directly on enterprise endpoints with system-level privileges. Learn how to detect and control AI agents in real time.
Interactive Challenge
AI Unlocked: Decoding Prompt Injection
You can’t secure AI until you know how to break it. Use your prompt injection skills to unlock secret phrases by outsmarting SAIGE, the AI chatbot guarding the system. The fewer tokens (words) you use, the higher your score.
AI Research
Taxonomy of Prompt Injection Methods
Prompt injection (PI) is the #1 OWASP risk for GenAI apps and LLMs, where attacker instructions trigger unintended or malicious behavior. CrowdStrike researchers track emerging PI methods, mapping how attacks reach LLMs and the techniques attackers use.
Highlights from the AI Summit:
Accelerating Secure AI Adoption and Development
AI Security Resources
Blog
CrowdStrike Research: Security Flaws in DeepSeek-Generated Code Linked to Political Triggers
