The Rise of Agentic Attacks: What They Look Like, and Where They're Headed

Wiz security researchers report an increasing trend of threat actors using AI agents to handle the hands-on work of cloud intrusions, including reconnaissance, exploitation, and post-exploitation. These agentic attacks operate at machine speed through automated harnesses, writing custom tooling and adapting iteratively while exhibiting distinct behavioral fingerprints such as variable cadences and self-documenting scripts. The authors note that while these attacks can scale rapidly and increase volume, their inherent noise and iterative mistakes also create new detection and defense opportunities. Open article →