Las Vegas, USA, July 29th, 2026, CyberNewswireCatches rogue AI agents – and stops them in live production before they cause damageSweet Security, the proactive runtime enforcement company for cloud and AI, today announced its further expansion into AI security with Agentic AI Blocking. Someone has to decide, in the moment, what an agent is allowed to do,” said Dror Kashti, CEO and co-founder of Sweet Security. “With most tools, you find out about bad behavior in a report after the damage is done. About Sweet SecuritySweet Security delivers proactive runtime enforcement for cloud and AI. Through the Sweet Learning Loop, organizations continuously Attack, Fix, and Defend, driving every application and AI agent toward maximal immunity.