Another 215papers do not have category tags yet and are not included in filtered results.
Proposes LGWM, using an action-conditioned world model to verify GUI agent screen transitions.
Measures citation fragility in AI search, demonstrating that ordinary public posts can enter citations and answers.
AbstractThe authors claim the paper quantifies vulnerabilities in AI search caused by citation preferences and low posting thresholds; tests show public web posts can penetrate AI citations, and they call for auditing the citation supply chain.
提出 MFOTL 离线时序验证以监控工具 Agent 的危险动作链
这篇论文对工具调用大语言模型智能体的安全监控进行了形式化时序验证研究。
提出 Tracekit，审计编码 Agent 的意图、推理与动作并检测账本篡改
Tracekit 是无第三方依赖的编码智能体审计层：把人类请求、模型自述和已执行动作并排写入可外部锚定的哈希链，执行前用正则门，事后用规则和独立模型互查。
提出 ORBIT 框架，评测多智能体防御对合谋等威胁的效果
AbstractORBIT 用六层配置比较多智能体防御，逐动作监控不迁移到合谋。
构建 ACE 语料，评测内核级证据对 Agent 攻击检测的有效性
ACE 是一份把智能体会话的内核 syscall 痕迹与应用层清单、转录配对起来的检测语料和基准。。
构建 COUNTER-GEO-BENCH，评测生成式搜索中的 GEO 虚假信息防御
Abstract论文构建了针对 GEO 虚假信息的首个防御基准，揭示通用护栏的失效与反作用，并提出了有效的轻量对比检测基线。
提出权威链劫持与轨迹引导策略演化，劫持搜索 Agent 的证据链
Abstract论文揭示了多步搜索智能体在受限中间人操纵下的脆弱性，提出 ACH 策略与 TGSE 演化方法并完成系统验证。
提出 S³，用 guard agent 编排阶段安全技能防御智能体多阶段风险
Proposes LAAF, which automatically attacks LPCI vulnerabilities in agent systems in stages
The authors state that LAAF is an automated red-teaming framework built specifically for LPCI in agentic LLMs. The issue to address is that encoded, delayed, and conditionally triggered payloads can persist in memory or RAG and re-enter across sessions, whereas Garak and PyRIT do not use the previous round's breakthrough results to seed the next stage across the six stages.
Proposes inference-time defense ICON, which detects attention over-focusing and redirects to correct indirect prompt injection in agents.
ICON detects attention over-focusing from IPI at inference time and redirects, avoiding refusal that would terminate the agent trajectory.
提出 Zombie Agent 攻击，经网页间接注入将载荷写入长期记忆并跨会话触发
Zombie Agent 把自进化 LLM 智能体的跨会话记忆当成攻击面：攻击者只发布网页，良性任务读入后，正常记忆更新留下载荷，并在以后无关会话里触发未授权工具行为。
用 OpenRAG-Soc 评测社交网页 RAG 的间接提示注入与检索投毒
提出 LPCI，把延迟条件载荷写入记忆与检索库以跨会话触发
作者把LPCI定义为针对智能体逻辑执行、记忆留存和工具集成的跨阶段漏洞类，而不是单次输入上的提示改写。
提出 SafeScientist 四层监控，阻止科研智能体执行高风险或不伦理任务
SafeScientist 是带四层监控的科研智能体框架，并用 SciSafetyBench 评测科学场景中的任务安全。