Memory Attack
No mentions found
This entity hasn't been tracked yet, or Iris is still building its knowledge base.
Related Articles from SNS
From Untrusted Input to Trusted Memory: A Systematic Study of Memory Poisoning Attacks in LLM Agents
arXiv:2606.04329v1 Announce Type: new Abstract: Memory is a core component of AI agents, enabling them to accumulate knowledge across interactions and improve performance. However, persistent memory introduces the risk of memory poisoning, where a single adversarial memory write can exert long-term influence over agent behavior. We present a systematic study of memory poisoning in LLM-based agents.
From Storage to Steering: Memory Control Flow Attacks on LLM Agents
Announce Type: replace Abstract: Modern agentic systems allow Large Language Model (LLM) agents to tackle complex tasks through extensive tool usage, forming structured control flows of tool selection and execution. Existing security analyses often treat these control flows as ephemeral, one-off sessions, overlooking the persistent influence of memory. This paper identifies a new threat from Memory Control Flow Attacks (MCFA) that memory can dominate the control flow, forcing unintended tool...
‘They took everything’: arson attack destroyed a mother’s memories of her dead son
Karen Holmes lost a son to cancer, then her home in Yorkshire to a fire; the house is now refurbished but its meaning has gone Karen Holmes is sitting in her newly renovated lounge in a house she has lived in for 28 years, but she cannot live here now. She cannot leave, either. The house looks good.
Topology Matters: Measuring Memory Leakage in Multi-Agent LLMs
arXiv:2512.04668v4 Announce Type: replace Abstract: Graph topology is a fundamental determinant of memory leakage in multi-agent LLM systems, yet its effects remain poorly quantified. We introduce MAMA (Multi-Agent Memory Attack), a controlled evaluation framework for comparing topology-conditioned memory leakage in multi-agent LLM systems. MAMA operates on synthetic documents containing labeled Personally Identifiable Information (PII) entities, from which we generate sanitized task...
Membrane: A Self-Evolving Contrastive Safety Memory for LLM Agent Defense
arXiv:2606.05743v1 Announce Type: new Abstract: Despite advances in safety alignment, large language models remain vulnerable to continuously evolving jailbreaks. Existing fine-tuned safety classifiers cannot adapt to these evolving attacks, while adaptive memory-based guardrails tend to over-refuse benign queries that resemble stored attacks. We propose Membrane, a self-evolving guardrail built on Contrastive Safety Memory (CSM): each cell pairs the conditions for blocking a harmful query...
Family's awful double heartache as man, 32, dies after younger brother fatally stabbed
Family's awful double heartache as man, 32, dies after younger brother fatally stabbed Thomas Griffin, a 32-year-old boxer, died after sustaining serious injuries in a suspected attack at the War Memorial Gardens by Islandbridge in west Dublin, with Gardaí now investigating the incident A young man has died after entering a river following a suspected slash attack. Thomas Griffin's death comes after his younger brother James was fatally stabbed at a club event. Police believe Thomas, 32, was...
Gaza Conflict: Two years on
To mark two years of the war in Gaza, our correspondents bring you this special episode of the Global News Podcast from the BBC Bureau in Jerusalem. Jon Donnison is joined by our Gaza correspondent Rushdi Abualouf, our International Editor Jeremy Bowen, and our Middle East correspondent Yolande Knell. They discuss the history of the conflict, their memories of the October 7th attacks, and respond to questions from BBC World Service listeners.
AI helped researchers bypass Apple M5 defenses
Apple devices have earned a reputation for being tough to break into. That comes from Apple's tight control over the hardware, software and many of the protections standing between you and an attacker. However, a new claim from security startup Calif shows how quickly the cybersecurity world may be changing.
AI helped researchers bypass Apple M5 defenses
Apple devices have earned a reputation for being tough to break into. That comes from Apple's tight control over the hardware, software and many of the protections standing between you and an attacker. However, a new claim from security startup Calif shows how quickly the cybersecurity world may be changing.
IND vs AFG: KL Rahul sheds T20 edginess, Sudharsan makes the most of 'long rope'
TimesofIndia.com in Mullanpur: The challenge before KL Rahul and Sai Sudharsan was not Afghanistan's bowling attack. It was the transition from two months of relentless T20 cricket back to the demands of Test cricket. One had to rediscover patience after a prolific IPL campaign, while the other had to justify the faith placed in him by the team management.