The Balkanization of Execution-Security Research for AI Coding Agents: Isolation, Access Control, and Time-of-Check-to-Time-of-Use Vulnerabilities
人工智能编码代理执行安全研究的巴尔干化:隔离、访问控制和检查时间到使用时间漏洞
机构 * Department of Computer Science(计算机科学系) ; AI and Media Analysis Lab(人工智能与媒体分析实验室) ; Berlin, Germany(德国柏林)
专题命中 软件智能体 :agent(abstract);agentic(abstract);分类 cs.AI
AI总结 研究人工智能编码代理执行安全相关文献分散问题,将39篇论文系统化分类并验证,找出五个交叉差距,如隔离架构与能力模型缺乏共享基准评估等,最后针对差距提出研究议程。
Comments 18 pages, 15 figures, 6 tables. Systematizes 39 execution-security papers (2023-2026) into 17 verified categories. Machine-readable corpus and verification script released as a supplementary artifact