AI Security
-
把安全告警交给 AI 之前：Defensive AI Agent 的工程化实践
从五类安全产品接入、证据归一化、产品 Agent、确定性验证、受控记忆到只读调查与双人审批，复盘 Defensive AI Agent 如何把模型能力放进一条可追溯、可恢复的 SOC 工作流。
-
Before Handing Security Alerts to AI: Engineering a Defensive AI Agent
A practical review of how a Defensive AI Agent turns heterogeneous security telemetry into a traceable, recoverable SOC workflow through evidence normalization, deterministic …