Persistent memory has enabled large language model (LLM) agents to store factual knowledge, prior decisions, reasoning histories, tool usage information, and context. While this has improved the agent's functionality and continuity across tasks, it has also introduced a new attack surface: the agent's own reasoning history. In this paper, we introduce the Forged Amplifying Rationale Memory Attack (FARMA), which poisons an agent's remembered reasoning rather than its factual knowledge. It inserts forged reasoning traces using evasive language that bypasses keyword-based defenses, then amplifies them through self-referential reinforcement that defeats consensus-based defenses. To address FARMA, we introduce SENTINEL, a layered defense pipeline to detect forged reasoning entries. Its central component is the Reasoning Guard that structurally analyzes candidate entries for forgery using five weighted signals.
We evaluate FARMA and SENTINEL across multiple agents and different LLM models with 50 trials and show that FARMA achieves an attack success rate of up to 100% under baseline conditions and is capable of defeating defense mechanisms like keyword filter and A-MemGuard. Our evaluation also shows that SENTINEL reduces FARMA's attack success rate to as low as 0% with no false positives observed across 326 benign agent traces. Our work demonstrates the need to protect not only an agent's retrieved content but also the integrity of its reasoning history.
MemAudit: Post-hoc Auditing of Poisoned Agent Memory via Causal Attribution and Structural Anomaly Detection
Institute of Information Engineering, Chinese Academy of Sciences(中国科学院信息工程研究所);Qiyuan Tech(齐元科技);Laboratory of Multimedia Information Processing, School of Computer Science, Peking University(北京大学计算机科学学院多媒体信息处理实验室);School of Cyber Security, University of Chinese Academy of Sciences(中国科学院大学网络安全学院)
EvoMail: Self-Evolving Cognitive Agents for Adaptive Spam and Phishing Email Defense
People’s Liberation Army 77606 Unit(中国人民解放军77606单位);Field Engineering College, Army Engineering University of PLA(陆军工程大学工程学院);College Of Software, ZhengZhou University of Light Industry(郑州轻工业大学软件学院)
State Key Laboratory of Integrated Services Networks (ISN)(综合业务网理论及关键技术国家重点实验室);Shaanxi Key Laboratory of Network and System Security, Xidian University(西安电子科技大学陕西省网络与系统安全重点实验室);School of Data Science and Engineering, East China Normal University(华东师范大学数据科学与工程学院);Ant Group(蚂蚁集团);College of Computer and Data Science, Fuzhou University(福州大学计算机与数据科学学院)
Selection Integrity for LLM Graph Memory: An Accumulability Criterion for Information-Flow-Blind Retrieval
University of Technology Sydney(悉尼科技大学);National University of Singapore(新加坡国立大学);University of Sheffield(谢菲尔德大学);University of New South Wales(新南威尔士大学)