Cyber attackers used a multi-agent AI framework to steal thousands of credentials in under six hours, a Google report said. Google Mandiant Threat Intelligence Group said a suspected financially motivated attacker breached an organisation\'s cloud infrastructure, built an autonomous attack system and automated scanning and information gathering. The attacker combined an AI coding chatbot with prompts and agent prompts, and used a preset markdown playbook. The report also cited a suspected China-linked group using Gemini for initial intrusion tasks.