ArborMem: Navigating Interaction States with Memory Forests
本文提出ArborMem，一种在线记忆框架，通过构建交互状态森林解决长对话中多任务、多人和计划交织时的记忆连续性问题。
0 citationsRead paper
Selected work
Representative Papers
SkillSentry: Adaptive Honey Worlds for Dynamic Safety Testing of Agent Skills
This work addresses the challenge that external skills invoked by large language model (LLM) agents may exhibit latent harmful behaviors under specific environmental conditions or interaction histories—risks that evade detection by existing static analysis methods. To tackle this, we propose SkillSentry, a dynamic security testing framework that simulates bait environments using LLMs, adaptively generates exploratory tasks, and compares execution trajectories with and without the target skill enabled. By correlating source code and runtime logs, SkillSentry enables precise attribution and detection of conditionally triggered malicious behaviors, overcoming the limitations of static approaches. Empirical evaluation shows that SkillSentry achieves 99.50% recall and an average F1 score of 96.26% on standard benchmarks. Notably, under semantic-preserving evasion attacks, it maintains robust performance with an average F1 of 92.95%, substantially outperforming the strongest baseline at 80.07%.
0 citationsRead paper
Dynamic Defense Profiling Enables Cognitive Jailbreak of Text-to-Image Models
Current jailbreaking attacks on text-to-image models suffer from low efficiency, semantic collapse, and neglect of critical information in defense feedback. This work proposes the MIND framework, which, for the first time, formulates jailbreaking as a cognitive reasoning process over latent defense mechanisms. By integrating multimodal feedback parsing, dynamically updating a defense profile, and employing meta-memory-driven strategy retrieval, MIND enables semantically coherent and adaptive attacks. The approach transcends the limitations of conventional black-box optimization, achieving a 95.62% attack success rate (ASR) across six defense configurations on Stable Diffusion v1.5 and up to 91.58% ASR on four major commercial text-to-image generation systems.
0 citationsRead paper
ElephantAgent: Contextual State Continuity in Agentic Systems
This work addresses the vulnerability of intelligent agent systems to context-state poisoning attacks—stemming from their reliance on external tools and memory—and the absence of verifiable guarantees for state continuity. To mitigate these risks, the authors propose ElephantAgent, a novel protocol that introduces state continuity mechanisms into dynamic context management for agent systems. By recomputing and verifying a digest of the local context state prior to each query and leveraging trusted hardware to maintain a linearized log of authorized state transitions, ElephantAgent ensures historical traceability. This approach effectively defends against attacks such as tool descriptor tampering and memory poisoning, while enabling anomaly detection and rollback to known-good states.
0 citationsRead paper
CL-CLIP: CLIP-Based Continual Learning Framework with Cost-Volume Category Decoupling for Object Detection
This work addresses catastrophic forgetting in continual object detection, where models tend to forget previously learned categories when acquiring new ones. To mitigate this issue, the authors propose a CLIP-based continual learning framework that incorporates a cost-volume-guided category decoupling mechanism. By leveraging CLIP’s image-text similarity to construct a cost volume, the method decomposes region features into category-specific pathways and introduces a multi-expert RoI detection head to jointly optimize learning of new classes and retention of old ones. The approach effectively integrates CLIP’s zero-shot priors into the continual learning process, substantially alleviating forgetting. Experimental results on PASCAL VOC and MS-COCO demonstrate that the proposed method outperforms existing baselines such as F-ViT, achieving state-of-the-art performance in both adapting to novel categories and preserving accuracy on base categories.
0 citationsRead paper
Recent publications
Latest Papers
ArborMem: Navigating Interaction States with Memory Forests
本文提出ArborMem，一种在线记忆框架，通过构建交互状态森林解决长对话中多任务、多人和计划交织时的记忆连续性问题。
0 citationsRead paper
SkillSentry: Adaptive Honey Worlds for Dynamic Safety Testing of Agent Skills
This work addresses the challenge that external skills invoked by large language model (LLM) agents may exhibit latent harmful behaviors under specific environmental conditions or interaction histories—risks that evade detection by existing static analysis methods. To tackle this, we propose SkillSentry, a dynamic security testing framework that simulates bait environments using LLMs, adaptively generates exploratory tasks, and compares execution trajectories with and without the target skill enabled. By correlating source code and runtime logs, SkillSentry enables precise attribution and detection of conditionally triggered malicious behaviors, overcoming the limitations of static approaches. Empirical evaluation shows that SkillSentry achieves 99.50% recall and an average F1 score of 96.26% on standard benchmarks. Notably, under semantic-preserving evasion attacks, it maintains robust performance with an average F1 of 92.95%, substantially outperforming the strongest baseline at 80.07%.
0 citationsRead paper
Dynamic Defense Profiling Enables Cognitive Jailbreak of Text-to-Image Models
Current jailbreaking attacks on text-to-image models suffer from low efficiency, semantic collapse, and neglect of critical information in defense feedback. This work proposes the MIND framework, which, for the first time, formulates jailbreaking as a cognitive reasoning process over latent defense mechanisms. By integrating multimodal feedback parsing, dynamically updating a defense profile, and employing meta-memory-driven strategy retrieval, MIND enables semantically coherent and adaptive attacks. The approach transcends the limitations of conventional black-box optimization, achieving a 95.62% attack success rate (ASR) across six defense configurations on Stable Diffusion v1.5 and up to 91.58% ASR on four major commercial text-to-image generation systems.
0 citationsRead paper
ElephantAgent: Contextual State Continuity in Agentic Systems
This work addresses the vulnerability of intelligent agent systems to context-state poisoning attacks—stemming from their reliance on external tools and memory—and the absence of verifiable guarantees for state continuity. To mitigate these risks, the authors propose ElephantAgent, a novel protocol that introduces state continuity mechanisms into dynamic context management for agent systems. By recomputing and verifying a digest of the local context state prior to each query and leveraging trusted hardware to maintain a linearized log of authorized state transitions, ElephantAgent ensures historical traceability. This approach effectively defends against attacks such as tool descriptor tampering and memory poisoning, while enabling anomaly detection and rollback to known-good states.
0 citationsRead paper
CL-CLIP: CLIP-Based Continual Learning Framework with Cost-Volume Category Decoupling for Object Detection
This work addresses catastrophic forgetting in continual object detection, where models tend to forget previously learned categories when acquiring new ones. To mitigate this issue, the authors propose a CLIP-based continual learning framework that incorporates a cost-volume-guided category decoupling mechanism. By leveraging CLIP’s image-text similarity to construct a cost volume, the method decomposes region features into category-specific pathways and introduces a multi-expert RoI detection head to jointly optimize learning of new classes and retention of old ones. The approach effectively integrates CLIP’s zero-shot priors into the continual learning process, substantially alleviating forgetting. Experimental results on PASCAL VOC and MS-COCO demonstrate that the proposed method outperforms existing baselines such as F-ViT, achieving state-of-the-art performance in both adapting to novel categories and preserving accuracy on base categories.
0 citationsRead paper