LLM agents increasingly rely on persistent, shared memory, turning the memory module into a new attack surface through which attackers can inject malicious text via query-only interaction. Existing query-only attacks, however, remain expensive, requiring multiple bridging injected queries per target and leaving visible directive traces in stored records. We present AMIA (Agent Memory Injection Attack), which improves both query efficiency and stealth through three injection families along a stealth-to-effectiveness axis: prompt injection, semi-instruction that hides the misleading answer in a benign restate or discredit request, and no-instruction that delivers the false claim as the presupposition of a natural follow-up question. We progressively scale AMIA from QA agents on flat key-value memory, to multi-agent shared memory, and to three production-oriented memory architectures (Mem0, A-MEM, MemoryOS), and finally to three non-QA agent classes (clinical support, code summarization, web shopping), where AMIA exceeds prior query-only attacks on most settings while using a fraction of their inject query budget. We further evaluate defenses that mitigate all three attack families across settings. Our code is available at: https://github.com/TrustAIRLab/AMIA.

Gossip has been investigated primarily through vignette-based surveys, which provide limited insights into the dynamic and adaptive gossip behaviors within real-world contexts. To address these limitations, we introduce an LLM agent-based simulation framework, designed as a configurable experimental laboratory for gossip research. We validate our framework across three core dimensions of gossip dynamics: initiation, reaction, and perception. Our framework can reproduce several behavioral trends observed in human studies and enable analysis of agents’ reasoning and temporally unfolding interactions. We further augment the framework with a chain-of-reasoning mechanism to support exploratory investigation, demonstrating how explicit agent reasoning and experiment-level controllability enable the generation of hypotheses about gossip dynamics beyond fixed survey settings. Overall, our framework bridges LLM-based computational modeling and social science, complementing vignette-based methodologies and enabling controlled, systematic study of complex social behaviors. Our code is available at: https://github.com/TrustAIRLab/Gossip-Simulacra.

Recently, autonomous agents built on large language models (LLMs) have experienced significant development and are being deployed in real-world applications. Through the usage of tools, these systems can perform actions in the real world. Given the agents’ practical applications and ability to execute consequential actions, such autonomous systems can cause more severe damage than a standalone LLM if compromised. While some existing research has explored harmful actions by LLM agents, our study approaches the vulnerability from a different perspective. We introduce a new type of attack that causes malfunctions by misleading the agent into executing repetitive or irrelevant actions. Our experiments reveal that these attacks can induce failure rates exceeding 80% in multiple scenarios. Through attacks on implemented and deployable agents in multi-agent scenarios, we accentuate the realistic risks associated with these vulnerabilities. To mitigate such attacks, we propose self-examination defense methods. Our findings indicate these attacks are more difficult to detect compared to previous overtly harmful attacks, highlighting the substantial risks associated with this vulnerability.