arXiv subjects

Zhiqiang Lin

Publications and source records attributed to Zhiqiang Lin.

At least 37 records · Page 2Linked to original sources

arXiv2026

Atomicity for Agents: Exposing, Exploiting, and Mitigating TOCTOU Vulnerabilities in Browser-Use Agents

Browser-use agents are widely used for everyday tasks. They enable automated interaction with web pages through structured DOM based interfaces or vision language models operating on page screenshots. However, web pages often change between planning and execution, causing agents to execute actions based on stale assumptions. We view this temporal mismatch as a time of check to time of use (TOCTOU) vulnerability in browser-use agents. Dynamic or adversarial web content can exploit this window to induce unintended actions. We present a large scale empirical study of TOCTOU vulnerabilities in browser-use agents using a benchmark that spans synthesized and real world websites. Using this benchmark, we evaluate 10 popular open source agents and show that TOCTOU vulnerabilities are widespread. We design a lightweight mitigation based on pre-execution validation. It monitors DOM and layout changes during planning and validates the page state immediately before action execution. This approach reduces the risk of insecure execution and mitigates unintended side effects in browser-use agents.

cs.CR↗

arXiv2026

SIN-Bench: Tracing Native Evidence Chains in Long-Context Multimodal Scientific Interleaved Literature

Evaluating whether multimodal large language models truly understand long-form scientific papers remains challenging: answer-only metrics and synthetic "Needle-In-A-Haystack" tests often reward answer matching without requiring a causal, evidence-linked reasoning trace in the document. We propose the "Fish-in-the-Ocean" (FITO) paradigm, which requires models to construct explicit cross-modal evidence chains within native scientific documents. To operationalize FITO, we build SIN-Data, a scientific interleaved corpus that preserves the native interleaving of text and figures. On top of it, we construct SIN-Bench with four progressive tasks covering evidence discovery (SIN-Find), hypothesis verification (SIN-Verify), grounded QA (SIN-QA), and evidence-anchored synthesis (SIN-Summary). We further introduce "No Evidence, No Score", scoring predictions when grounded to verifiable anchors and diagnosing evidence quality via matching, relevance, and logic. Experiments on eight MLLMs show that grounding is the primary bottleneck: Gemini-3-pro achieves the best average overall score (0.573), while GPT-5 attains the highest SIN-QA answer accuracy (0.767) but underperforms on evidence-aligned overall scores, exposing a gap between correctness and traceable support.

cs.CL↗

arXiv2025

AnyCap Project: A Unified Framework, Dataset, and Benchmark for Controllable Omni-modal Captioning

Controllable captioning is essential for precise multimodal alignment and instruction following, yet existing models often lack fine-grained control and reliable evaluation protocols. To address this gap, we present the AnyCap Project, an integrated solution spanning model, dataset, and evaluation. We introduce AnyCapModel (ACM), a lightweight plug-and-play framework that enhances the controllability of existing foundation models for omni-modal captioning without retraining the base model. ACM reuses the original captions from base models while incorporating user instructions and modality features to generate improved captions. To remedy the data scarcity in controllable multimodal captioning, we build AnyCapDataset (ACD), covering three modalities, 28 user-instruction types, and 300\,k high-quality data entries. We further propose AnyCapEval, a new benchmark that provides more reliable evaluation metrics for controllable captioning by decoupling content accuracy and stylistic fidelity. ACM markedly improves caption quality across a diverse set of base models on AnyCapEval. Notably, ACM-8B raises GPT-4oś content scores by 45\% and style scores by 12\%, and it also achieves substantial gains on widely used benchmarks such as MIA-Bench and VidCapBench.

cs.CV↗

arXiv2025

MobiLLM: An Agentic AI Framework for Closed-Loop Threat Mitigation in 6G Open RANs

The evolution toward 6G networks is being accelerated by the Open Radio Access Network (O-RAN) paradigm -- an open, interoperable architecture that enables intelligent, modular applications across public telecom and private enterprise domains. While this openness creates unprecedented opportunities for innovation, it also expands the attack surface, demanding resilient, low-cost, and autonomous security solutions. Legacy defenses remain largely reactive, labor-intensive, and inadequate for the scale and complexity of next-generation systems. Current O-RAN applications focus mainly on network optimization or passive threat detection, with limited capability for closed-loop, automated response. To address this critical gap, we present an agentic AI framework for fully automated, end-to-end threat mitigation in 6G O-RAN environments. MobiLLM orchestrates security workflows through a modular multi-agent system powered by Large Language Models (LLMs). The framework features a Threat Analysis Agent for real-time data triage, a Threat Classification Agent that uses Retrieval-Augmented Generation (RAG) to map anomalies to specific countermeasures, and a Threat Response Agent that safely operationalizes mitigation actions via O-RAN control interfaces. Grounded in trusted knowledge bases such as the MITRE FiGHT framework and 3GPP specifications, and equipped with robust safety guardrails, MobiLLM provides a blueprint for trustworthy AI-driven network security. Initial evaluations demonstrate that MobiLLM can effectively identify and orchestrate complex mitigation strategies, significantly reducing response latency and showcasing the feasibility of autonomous security operations in 6G.

cs.CR↗

arXiv2025

A-MemGuard: A Proactive Defense Framework for LLM-Based Agent Memory

Large Language Model (LLM) agents use memory to learn from past interactions, enabling autonomous planning and decision-making in complex environments. However, this reliance on memory introduces a critical security risk: an adversary can inject seemingly harmless records into an agent's memory to manipulate its future behavior. This vulnerability is characterized by two core aspects: First, the malicious effect of injected records is only activated within a specific context, making them hard to detect when individual memory entries are audited in isolation. Second, once triggered, the manipulation can initiate a self-reinforcing error cycle: the corrupted outcome is stored as precedent, which not only amplifies the initial error but also progressively lowers the threshold for similar attacks in the future. To address these challenges, we introduce A-MemGuard (Agent-Memory Guard), the first proactive defense framework for LLM agent memory. The core idea of our work is the insight that memory itself must become both self-checking and self-correcting. Without modifying the agent's core architecture, A-MemGuard combines two mechanisms: (1) consensus-based validation, which detects anomalies by comparing reasoning paths derived from multiple related memories and (2) a dual-memory structure, where detected failures are distilled into ``lessons'' stored separately and consulted before future actions, breaking error cycles and enabling adaptation. Comprehensive evaluations on multiple benchmarks show that A-MemGuard effectively cuts attack success rates by over 95% while incurring a minimal utility cost. This work shifts LLM memory security from static filtering to a proactive, experience-driven model where defenses strengthen over time. Our code is available in https://github.com/TangciuYueng/AMemGuard

cs.CR↗

arXiv2025

PIN: A Knowledge-Intensive Dataset for Paired and Interleaved Multimodal Documents

Recent advancements in large multimodal models (LMMs) have leveraged extensive multimodal datasets to enhance capabilities in complex knowledge-driven tasks. However, persistent challenges in perceptual and reasoning errors limit their efficacy, particularly in interpreting intricate visual data and deducing multimodal relationships. To address these issues, we introduce PIN (Paired and INterleaved multimodal documents), a novel data format designed to foster a deeper integration of visual and textual knowledge. The PIN format uniquely combines semantically rich Markdown files, which preserve fine-grained textual structures, with holistic overall images that capture the complete document layout. Following this format, we construct and release two large-scale, open-source datasets: PIN-200M (~200 million documents) and PIN-14M (~14 million), compiled from diverse web and scientific sources in both English and Chinese. To maximize usability, we provide detailed statistical analyses and equip the datasets with quality signals, enabling researchers to easily filter and select data for specific tasks. Our work provides the community with a versatile data format and substantial resources, offering a foundation for new research in pre-training strategies and the development of more powerful knowledge-intensive LMMs.

cs.AI↗

arXiv2025

AUTOVR: Automated UI Exploration for Detecting Sensitive Data Flow Exposures in Virtual Reality Apps

The rise of Virtual Reality (VR) has provided developers with an unprecedented platform for creating games and applications (apps) that require distinct inputs, different from those of conventional devices like smartphones. The Meta Quest VR platform, driven by Meta, has democratized VR app publishing and attracted millions of users worldwide. However, as the number of published apps grows, there is a notable lack of robust headless tools for user interface (UI) exploration and user event testing. To address this need, we present AUTOVR, an automatic framework for dynamic UI and user event interaction in VR apps built on the Unity Engine. Unlike conventional Android and GUI testers, AUTOVR analyzes the app's internal binary to reveal hidden events, resolves generative event dependencies, and utilizes them for comprehensive exploration of VR apps. Using sensitive data exposure as a performance metric, we compare AUTOVR with Android Monkey, a widely used headless Android GUI stress testing tool. Our empirical evaluation demonstrates AUTOVR's superior performance, triggering an order of magnitude of more sensitive data exposures and significantly enhancing the privacy of VR apps.

cs.CR↗

arXiv2025

ELFuzz: Efficient Input Generation via LLM-driven Synthesis Over Fuzzer Space

Generation-based fuzzing produces appropriate test cases according to specifications of input grammars and semantic constraints to test systems and software. However, these specifications require significant manual effort to construct. This paper proposes a new approach, ELFuzz (Evolution Through Large Language Models for Fuzzing), that automatically synthesizes generation-based fuzzers tailored to a system under test (SUT) via LLM-driven synthesis over fuzzer space. At a high level, it starts with minimal seed fuzzers and propels the synthesis by fully automated LLM-driven evolution with coverage guidance. Compared to previous approaches, ELFuzz can 1) seamlessly scale to SUTs of real-world sizes -- up to 1,791,104 lines of code in our evaluation -- and 2) synthesize efficient fuzzers that catch interesting grammatical structures and semantic constraints in a human-understandable way. Our evaluation compared ELFuzz with specifications manually written by domain experts and synthesized by state-of-the-art approaches. It shows that ELFuzz achieves up to 434.8% more coverage over the second best and triggers up to 216.7% more artificially injected bugs, compared to the state-of-the-art. We also used ELFuzz to conduct a real-world fuzzing campaign on the newest version of cvc5 for 14 days, and encouragingly, it found five 0-day bugs (three are exploitable). Moreover, we conducted an ablation study, which shows that the fuzzer space model, the key component of ELFuzz, contributes the most (up to 62.5%) to the effectiveness of ELFuzz. Further analysis of the fuzzers synthesized by ELFuzz confirms that they catch interesting grammatical structures and semantic constraints in a human-understandable way. The results present the promising potential of ELFuzz for more automated, efficient, and extensible input generation for fuzzing.

cs.CR↗

arXiv2025

ARMOUR US: Android Runtime Zero-permission Sensor Usage Monitoring from User Space

This work investigates how to monitor access to Android zero-permission sensors which could cause privacy leakage to users. Moreover, monitoring such sensitive access allows security researchers to characterize potential sensor abuse patterns. Zero-permission sensors such as accelerometers have become an indispensable part of Android devices. The critical information they provide has attracted extensive research investigating how data collectors could capture more sensor data to enable both benign and exploitative applications. In contrast, little work has explored how to enable data providers, such as end users, to understand sensor usage. While existing methods such as static analysis and hooking-based dynamic analysis face challenges of requiring complicated development chains, rooting privilege, and app-specific reverse engineering analysis, our work aims to bridge this gap by developing ARMOUR for user-space runtime monitoring, leveraging the intrinsic sampling rate variation and convergence behaviors of Android. ARMOUR enables privacy-aware users to easily monitor how third-party apps use sensor data and support security researchers to perform rapid app-agnostic sensor access analysis. Our evaluation with 1,448 commercial applications shows the effectiveness of ARMOUR in detecting sensor usage in obfuscated code and other conditions, and observes salient sensor abuse patterns such as 50% of apps from seemingly sensor-independent categories accessing data of multiple zero-permission sensors. We analyze the impact of Android's recent policy changes on zero-permission sensors and remaining technical and regulatory problems.

cs.CR↗

arXiv2025

JsDeObsBench: Measuring and Benchmarking LLMs for JavaScript Deobfuscation

Deobfuscating JavaScript (JS) code poses a significant challenge in web security, particularly as obfuscation techniques are frequently used to conceal malicious activities within scripts. While Large Language Models (LLMs) have recently shown promise in automating the deobfuscation process, transforming detection and mitigation strategies against these obfuscated threats, a systematic benchmark to quantify their effectiveness and limitations has been notably absent. To address this gap, we present JsDeObsBench, a dedicated benchmark designed to rigorously evaluate the effectiveness of LLMs in the context of JS deobfuscation. We detail our benchmarking methodology, which includes a wide range of obfuscation techniques ranging from basic variable renaming to sophisticated structure transformations, providing a robust framework for assessing LLM performance in real-world scenarios. Our extensive experimental analysis investigates the proficiency of cutting-edge LLMs, e.g., GPT-4o, Mixtral, Llama, and DeepSeek-Coder, revealing superior performance in code simplification despite challenges in maintaining syntax accuracy and execution reliability compared to baseline methods. We further evaluate the deobfuscation of JS malware to exhibit the potential of LLMs in security scenarios. The findings highlight the utility of LLMs in deobfuscation applications and pinpoint crucial areas for further improvement.

cs.CR↗

arXiv2025

AI Safety vs. AI Security: Demystifying the Distinction and Boundaries

Artificial Intelligence (AI) is rapidly being integrated into critical systems across various domains, from healthcare to autonomous vehicles. While its integration brings immense benefits, it also introduces significant risks, including those arising from AI misuse. Within the discourse on managing these risks, the terms "AI Safety" and "AI Security" are often used, sometimes interchangeably, resulting in conceptual confusion. This paper aims to demystify the distinction and delineate the precise research boundaries between AI Safety and AI Security. We provide rigorous definitions, outline their respective research focuses, and explore their interdependency, including how security breaches can precipitate safety failures and vice versa. Using clear analogies from message transmission and building construction, we illustrate these distinctions. Clarifying these boundaries is crucial for guiding precise research directions, fostering effective cross-disciplinary collaboration, enhancing policy effectiveness, and ultimately, promoting the deployment of trustworthy AI systems.

cs.CY↗

arXiv2025

C2RUST-BENCH: A Minimized, Representative Dataset for C-to-Rust Transpilation Evaluation

Despite the effort in vulnerability detection over the last two decades, memory safety vulnerabilities continue to be a critical problem. Recent reports suggest that the key solution is to migrate to memory-safe languages. To this end, C-to-Rust transpilation becomes popular to resolve memory-safety issues in C programs. Recent works propose C-to-Rust transpilation frameworks; however, a comprehensive evaluation dataset is missing. Although one solution is to put together a large enough dataset, this increases the analysis time in automated frameworks as well as in manual efforts for some cases. In this work, we build a method to select functions from a large set to construct a minimized yet representative dataset to evaluate the C-to-Rust transpilation. We propose C2RUST-BENCH that contains 2,905 functions, which are representative of C-to-Rust transpilation, selected from 15,503 functions of real-world programs.

cs.CR↗

arXiv2025

Demystifying Private Transactions and Their Impact in PoW and PoS Ethereum

In Ethereum, private transactions, a specialized transaction type employed to evade public Peer-to-Peer (P2P) network broadcasting, remain largely unexplored, particularly in the context of the transition from Proof-of-Work (PoW) to Proof-of-Stake (PoS) consensus mechanisms. To address this gap, we investigate the transaction characteristics, (un)intended usages, and monetary impacts by analyzing large-scale datasets comprising 14,810,392 private transactions within a 15.5-month PoW dataset and 30,062,232 private transactions within a 15.5-month PoS dataset. While originally designed for security purposes, we find that private transactions predominantly serve three distinct functions in both PoW and PoS Ethereum: extracting Maximum Extractable Value (MEV), facilitating monetary transfers to distribute mining rewards, and interacting with popular Decentralized Finance (DeFi) applications. Furthermore, we find that private transactions are utilized in DeFi attacks to circumvent surveillance by white hat monitors, with an increased prevalence observed in PoS Ethereum compared to PoW Ethereum. Additionally, in PoS Ethereum, there is a subtle uptick in the role of private transactions for MEV extraction. This shift could be attributed to the decrease in transaction costs. However, this reduction in transaction cost and the cancellation of block rewards result in a significant decrease in mining profits for block creators.

cs.CR↗

arXiv2025

Ultraverse: A System-Centric Framework for Efficient What-If Analysis for Database-Intensive Web Applications

Existing what-if analysis systems are predominantly tailored to operate on either only the application layer or only the database layer of software. This isolated approach limits their effectiveness in scenarios where intensive interaction between applications and database systems occurs. To address this gap, we introduce Ultraverse, a what-if analysis framework that seamlessly integrates both application and database layers. Ultraverse employs dynamic symbolic execution to effectively translate application code into compact SQL procedure representations, thereby synchronizing application semantics at both SQL and application levels during what-if replays. A novel aspect of Ultraverse is its use of advanced query dependency analysis, which serves two key purposes: (1) it eliminates the need to replay irrelevant transactions that do not influence the outcome, and (2) it facilitates parallel replay of mutually independent transactions, significantly enhancing the analysis efficiency. Ultraverse is applicable to existing unmodified database systems and legacy application codes. Our extensive evaluations of the framework have demonstrated remarkable improvements in what-if analysis speed, achieving performance gains ranging from 7.7x to 291x across diverse benchmarks.

cs.DB↗

arXiv2024

Ditto: Elastic Confidential VMs with Secure and Dynamic CPU Scaling

Confidential Virtual Machines (CVMs) are a type of VMbased Trusted Execution Environments (TEEs) designed to enhance the security of cloud-based VMs, safeguarding them even from malicious hypervisors. Although CVMs have been widely adopted by major cloud service providers, current CVM designs face significant challenges in runtime resource management due to their fixed capacities and lack of transparency. These limitations hamper efficient cloud resource management, leading to increased operational costs and reduced agility in responding to fluctuating workloads. This paper introduces a dynamic CPU resource management approach, featuring the novel concept of "Elastic CVM. This approach allows for hypervisor-assisted runtime adjustment of CPU resources using a specialized vCPU type, termed Worker vCPU. This new approach enhances CPU resource adaptability and operational efficiency without compromising security. Additionally, we introduce a Worker vCPU Abstraction Layer to simplify Worker vCPU deployment and management. To demonstrate the effectiveness of our approach, we have designed and implemented a serverless computing prototype platform, called Ditto. We show that Ditto significantly improves performance and efficiency through finergrain resource management. The concept of "Elastic CVM" and the Worker vCPU design not only optimize cloud resource utilization but also pave the way for more flexible and cost-effective confidential computing environments.

cs.CR↗

arXiv2024

BaseMirror: Automatic Reverse Engineering of Baseband Commands from Android's Radio Interface Layer

In modern mobile devices, baseband is an integral component running on top of cellular processors to handle crucial radio communications. However, recent research reveals significant vulnerabilities in these basebands, posing serious security risks like remote code execution. Yet, effectively scrutinizing basebands remains a daunting task, as they run closed-source and proprietary software on vendor-specific chipsets. Existing analysis methods are limited by their dependence on manual processes and heuristic approaches, reducing their scalability. This paper introduces a novel approach to unveil security issues in basebands from a unique perspective: to uncover vendor-specific baseband commands from the Radio Interface Layer (RIL), a hardware abstraction layer interfacing with basebands. To demonstrate this concept, we have designed and developed BaseMirror, a static binary analysis tool to automatically reverse engineer baseband commands from vendor-specific RIL binaries. It utilizes a bidirectional taint analysis algorithm to adeptly identify baseband commands from an enhanced control flow graph enriched with reconstructed virtual function calls. Our methodology has been applied to 28 vendor RIL libraries, encompassing a wide range of Samsung Exynos smartphone models on the market. Remarkably, BaseMirror has uncovered 873 unique baseband commands undisclosed to the public. Based on these results, we develop an automated attack discovery framework to successfully derive and validate 8 zero-day vulnerabilities that trigger denial of cellular service and arbitrary file access on a Samsung Galaxy A53 device. These findings have been reported and confirmed by Samsung and a bug bounty was awarded to us.

cs.CR↗

arXiv2024

Flashback: Enhancing Proposer-Builder Design with Future-Block Auctions in Proof-of-Stake Ethereum

Maximal extractable value (MEV) in which block proposers unethically gain profits by manipulating the order in which transactions are included within a block, is a key challenge facing blockchains such as Ethereum today. Left unchecked, MEV can lead to a centralization of stake distribution thereby ultimately compromising the security of blockchain consensus. To preserve proposer decentralization (and hence security) of the blockchain, Ethereum has advocated for a proposer-builder separation (PBS) in which the functionality of transaction ordering is separated from proposers and assigned to separate entities called builders. Builders accept transaction bundles from searchers, who compete to find the most profitable bundles. Builders then bid completed blocks to proposers, who accept the most profitable blocks for publication. The auction mechanisms used between searchers, builders and proposers are crucial to the overall health of the blockchain. In this paper, we consider PBS design in Ethereum as a game between searchers, builders and proposers. A key novelty in our design is the inclusion of future block proposers, as all proposers of an epoch are decided ahead of time in proof-of-stake (PoS) Ethereum within the game model. Our analysis shows the existence of alternative auction mechanisms that result in a better (more profitable) equilibrium to players compared to state-of-the-art. Experimental evaluations based on synthetic and real-world data traces corroborate the analysis. Our results highlight that a rethinking of auction mechanism designs is necessary in PoS Ethereum to prevent disruption.

cs.CR↗

arXiv2024

SoK: Where's the "up"?! A Comprehensive (bottom-up) Study on the Security of Arm Cortex-M Systems

Arm Cortex-M processors are the most widely used 32-bit microcontrollers among embedded and Internet-of-Things devices. Despite the widespread usage, there has been little effort in summarizing their hardware security features, characterizing the limitations and vulnerabilities of their hardware and software stack, and systematizing the research on securing these systems. The goals and contributions of this paper are multi-fold. First, we analyze the hardware security limitations and issues of Cortex-M systems. Second, we conducted a deep study of the software stack designed for Cortex-M and revealed its limitations, which is accompanied by an empirical analysis of 1,797 real-world firmware. Third, we categorize the reported bugs in Cortex-M software systems. Finally, we systematize the efforts that aim at securing Cortex-M systems and evaluate them in terms of the protections they offer, runtime performance, required hardware features, etc. Based on the insights, we develop a set of recommendations for the research community and MCU software developers.

cs.CR↗