1 article on this topic.
The new OWASP GenAI/LLM Top 10 blends expert consensus with 6,639 documented real-world incidents — and the shift shows agentic AI deployments are already getting breached in production.