A sophisticated cyberattack on a Taiwan government network was carried out by China-linked hackers using eight autonomous AI agents, which breached the system, stole data, and compromised accounts with minimal human intervention. The attack, documented by Israeli cybersecurity firm Dream, occurred over four days in early July and involved a tool built from publicly available AI agents that mapped 21 government systems, identified vulnerabilities, and adapted its tactics as needed. This operation is notable for being the first fully autonomous, end-to-end AI hacking operation against a government target, demonstrating a significant evolution in attack methods. The use of AI agents allowed the hackers to operate with increased speed and efficiency, highlighting the potential for similar attacks in the future1. This development matters to cybersecurity practitioners because it signals a new era of autonomous attacks, which will require adaptive defensive strategies to counter.