13 papers
MMSkillRisk: Can Agents Stay Safe When Multimodal Skills Become Traps?
Lingqi Jiang, Jialuo Chen, Jianan Ma +8
Agent skills are shareable packages of procedural instructions, tools, and examples. Multimodal skills additionally include visual references that agents retrieve and inspect durin…
Safety in Self-Evolving LLM Agent Systems: Threats, Amplification, and Case Studies
Ruixiao Lin, Xinhao Deng, Qingming Li +12
Self-evolving LLM agent systems, which autonomously update their model parameters, memory, tools, and architectures, introduce a qualitatively new threat landscape in which adversa…
Constitutional On-Policy Safe Distillation
Ming Wen, Yuxuan Liu, Kun Yang +8
On-policy self-distillation (OPSD) has emerged as an efficient post-training paradigm by using a teacher conditioned on privileged information to provide dense token-level supervis…
ASEval: Automated Trajectory-Level Security Testing for Autonomous Agents
Jianan Ma, Xiaohu Du, Ruixiao Lin +9
As autonomous agents (e.g., OpenClaw) increasingly operate with deep system-level privileges to execute complex tasks, they introduce severe, unmitigated security risks. Existing L…
Taming OpenClaw: Security Analysis and Mitigation of Autonomous LLM Agent Threats
Xinhao Deng, Yixiang Zhang, Jiaqing Wu +15
Autonomous Large Language Model (LLM) agents, exemplified by OpenClaw, demonstrate remarkable capabilities in executing complex, long-horizon tasks. However, their tightly coupled…
OOD-MMSafe: Advancing MLLM Safety from Harmful Intent to Hidden Consequences
Ming Wen, Kun Yang, Jingyu Zhang +4
While safety alignment for Multimodal Large Language Models (MLLMs) has gained significant attention, current paradigms primarily target malicious intent or situational violations.…
Pragma-VL: Towards a Pragmatic Arbitration of Safety and Helpfulness in MLLMs
Ming Wen, Kun Yang, Xin Chen +4
Multimodal Large Language Models (MLLMs) pose critical safety challenges, as they are susceptible not only to adversarial attacks such as jailbreaking but also to inadvertently gen…
The Reasoning Trap: How Enhancing LLM Reasoning Amplifies Tool Hallucination
Chenlong Yin, Zeyang Sha, Shiwen Cui +2
Enhancing the reasoning capabilities of Large Language Models (LLMs) is a key strategy for building Agents that "think then act." However, recent observations, like OpenAI's o3, su…
Agent Safety Alignment via Reinforcement Learning
Zeyang Sha, Hanling Tian, Zhuoer Xu +3
The emergence of autonomous Large Language Model (LLM) agents capable of tool usage has introduced new safety risks that go beyond traditional conversational misuse. These agents,…
SEM: Reinforcement Learning for Search-Efficient Large Language Models
Zeyang Sha, Shiwen Cui, Weiqiang Wang
Recent advancements in Large Language Models(LLMs) have demonstrated their capabilities not only in reasoning but also in invoking external tools, particularly search engines. Howe…
Can VLMs Detect and Localize Fine-Grained AI-Edited Images?
Zhen Sun, Ziyi Zhang, Zeren Luo +10
Fine-grained detection and localization of localized image edits is crucial for assessing content authenticity, especially as modern diffusion models and image editors can produce…
GenPTW: Latent Image Watermarking for Provenance Tracing and Tamper Localization
Zhenliang Gan, Chunya Liu, Yichao Tang +4
The proliferation of generative image models has revolutionized AIGC creation while amplifying concerns over content provenance and manipulation forensics. Existing methods are typ…
Thought Manipulation: External Thought Can Be Efficient for Large Reasoning Models
Yule Liu, Jingyi Zheng, Zhen Sun +6
Recent advancements in large reasoning models (LRMs) have demonstrated the effectiveness of scaling test-time computation to enhance reasoning capabilities on various tasks. Howeve…