ICML

Cross-Agent Campaign Attribution: Linking Asynchronous Attacks Across LLM Agents

跨智能体活动归因：关联大语言模型智能体间的异步攻击

SangJin Park, Myungsub Choi, Jineok Kim, Minseung Kang

AI总结 研究跨大语言模型智能体的异步攻击归因问题，提出异步归因指纹向量（$A^2FV$）协议，构建SCD-v1基准，实验表明$A^2FV$在活动关联上表现良好，确立跨智能体活动归因作为保护大语言模型智能体的独特评估层。

Comments 22 pages, 5 figures. Accepted at the Second Workshop on Agents in the Wild: Safety, Security, and Beyond (AIWILD) at ICML 2026