Sweet Security Blocks Rogue AI Agents in Real Time

Sweet Security launched Agentic AI Blocking, stopping unauthorized AI agent actions in production. The feature terminates tool calls, blocks data leaks, and shuts down prompt injections. It analyzes over one billion runtime events daily. Early customer Zoomd uses the platform. Sweet will demonstrate at Black Hat USA 2026.