Prompt Injection Is No Longer the Main Problem
中文摘要
AI智能体具备行动能力后，安全重点将从提示词注入转向操作系统和云工作负载安全。
English Summary
As AI agents gain action capabilities, security must shift from prompt injection to operating-system and cloud workload security.
Original Excerpt
Once agents can act, security has to look more like operating-system and cloud workload security. Continue reading on Medium »