Haoran Ou

Xuanye Zhang, Yongsen Zheng, Zhuqin Xu et al.

LLM-driven agents are capable of selecting external tools to complete users' tasks. However, attackers could compromise such process, steering agents toward inappropriate/wrong tools and enabling malicious actions. Most existing attacks primarily manipulate the tool metadata, which is easily detectable by auditing and may lose effectiveness as modern agents increasingly adopt memory modules to refine tool selection policies through accumulated experience. This paper proposes MemMorph, the first attack that bias tool selection by poisoning the agent's long-term memory. Rather than explicitly dictating the tool invocation decision, MemMorph injects a small number of crafted records that are disguised as technical facts, incident reports, and operational policies. These poisoned records reshape the agent's contextual perception and decision-making process, leading it to autonomously infer and select the tool preferred by the attacker. Experiments across 3 benchmarks, 10 agent backbones, and 3 memory-module implementations show that MemMorph achieves up to 85.9% attack success rate with only three injected records, outperforming the strongest baseline by up to 25% while retaining potency under 3 representative defenses. Our findings expose long-term memory as a critical and under-explored attack surface in tool-augmented agents, urging the development of memory-level integrity safeguards.

Haoran Ou, Kangjie Chen, Gelei Deng et al.

Fact-checking systems with search-enabled large language models (LLMs) have shown strong potential for verifying claims by dynamically retrieving external evidence. However, the robustness of such systems against adversarial attack remains insufficiently understood. In this work, we study adversarial claim attacks against search-enabled LLM-based fact-checking systems under a realistic input-only threat model. We propose DECEIVE-AFC, an agent-based adversarial attack framework that integrates novel claim-level attack strategies and adversarial claim validity evaluation principles. DECEIVE-AFC systematically explores adversarial attack trajectories that disrupt search behavior, evidence retrieval, and LLM-based reasoning without relying on access to evidence sources or model internals. Extensive evaluations on benchmark datasets and real-world systems demonstrate that our attacks substantially degrade verification performance, reducing accuracy from 78.7% to 53.7%, and significantly outperform existing claim-based attack baselines with strong cross-system transferability.

Jiaren Peng, Zeqin Li, Chang You et al.

The rapid advancement of Large Language Models (LLMs) has created new opportunities for Automated Penetration Testing (AutoPT), spawning numerous frameworks aimed at achieving end-to-end autonomous attacks. However, despite the proliferation of related studies, existing research generally lacks systematic architectural analysis and large-scale empirical comparisons under a unified benchmark. Therefore, this paper presents the first Systematization of Knowledge (SoK) focusing on the architectural design and comprehensive empirical evaluation of current LLM-based AutoPT frameworks. At systematization level, we comprehensively review existing framework designs across six dimensions: agent architecture, agent plan, agent memory, agent execution, external knowledge, and benchmarks. At empirical level, we conduct large-scale experiments on 13 representative open-source AutoPT frameworks and 2 baseline frameworks utilizing a unified benchmark. The experiments consumed over 10 billion tokens in total and generated more than 1,500 execution logs, which were manually reviewed and analyzed over four months by a panel of more than 15 researchers with expertise in cybersecurity. By investigating the latest progress in this rapidly developing field, we provide researchers with a structured taxonomy to understand existing LLM-based AutoPT frameworks and a large-scale empirical benchmark, along with promising directions for future research.

Haoran Ou, Gelei Deng, Xingshuo Han et al.

The rise of Internet connectivity has accelerated the spread of disinformation, threatening societal trust, decision-making, and national security. Disinformation has evolved from simple text to complex multimodal forms combining images and text, challenging existing detection methods. Traditional deep learning models struggle to capture the complexity of multimodal disinformation. Inspired by advances in AI, this study explores using Large Language Models (LLMs) for automated disinformation detection. The empirical study shows that (1) LLMs alone cannot reliably assess the truthfulness of claims; (2) providing relevant evidence significantly improves their performance; (3) however, LLMs cannot autonomously search for accurate evidence. To address this, we propose Holmes, an end-to-end framework featuring a novel evidence retrieval method that assists LLMs in collecting high-quality evidence. Our approach uses (1) LLM-powered summarization to extract key information from open sources and (2) a new algorithm and metrics to evaluate evidence quality. Holmes enables LLMs to verify claims and generate justifications effectively. Experiments show Holmes achieves 88.3% accuracy on two open-source datasets and 90.2% in real-time verification tasks. Notably, our improved evidence retrieval boosts fact-checking accuracy by 30.8% over existing methods

Haoran Ou, Kangjie Chen, Xingshuo Han et al.

Large Language Models (LLMs) excel at tasks such as dialogue, summarization, and question answering, yet they struggle to adapt to specialized domains and evolving facts. To overcome this, web search has been integrated into LLMs, allowing real-time access to online content. However, this connection magnifies safety risks, as adversarial prompts combined with untrusted sources can cause severe vulnerabilities. We investigate red teaming for LLMs with web search and present CREST-Search, a framework that systematically exposes risks in such systems. Unlike existing methods for standalone LLMs, CREST-Search addresses the complex workflow of search-enabled models by generating adversarial queries with in-context learning and refining them through iterative feedback. We further construct WebSearch-Harm, a search-specific dataset to fine-tune LLMs into efficient red-teaming agents. Experiments show that CREST-Search effectively bypasses safety filters and reveals vulnerabilities in modern web-augmented LLMs, underscoring the need for specialized defenses to ensure trustworthy deployment.

Hangcheng Liu, Xu Kuang, Xingshuo Han et al.

Stereo Depth Estimation (SDE) is essential for scene perception in vision-based systems such as autonomous driving. Prior work shows SDE is vulnerable to pixel-optimization attacks, but these methods are limited to digital, static, and view-specific settings, making them impractical. This raises a central question: how to design deployable, adaptive, and transferable attacks under realistic constraints? We present two contributions to answer it. First, we build a unified framework that extends pixel-optimization attacks to four stereo-matching stages: feature extraction, cost-volume construction, cost aggregation, and disparity regression. Through systematic evaluation across nine SDE models with realistic constraints like photometric consistency, we show existing attacks suffer from poor transferability. Second, we propose PatchHunter, the first pixel-optimization-free attack. PatchHunter casts patch generation as a search in a structured space of visual patterns that disrupt core SDE assumptions, and uses a reinforcement learning policy to discover effective and transferable patterns efficiently. We evaluate PatchHunter on three levels: autonomous driving dataset, high-fidelity simulator, and real-world deployment. On KITTI, PatchHunter outperforms pixel-level attacks in both effectiveness and black-box transferability. Tests in CARLA and on vehicles with industrial-grade stereo cameras confirm robustness to physical variations. Even under challenging conditions such as low lighting, PatchHunter achieves a D1-all error above 0.4, while pixel-level attacks remain near 0.