Existing fuzzing tools struggle to analyze proprietary Industrial Communication Protocols (ICPs) due to the lack of detailed protocol specifications. Consequently, they often rely on manual analysis or coarse-grained black-box approaches, which offer limited insight into protocol field semantics. This ultimately restricts both the coverage and depth of the testing process. To address these limitations, we propose a novel fuzzing framework called ICProFuzz, which leverages Large Language Models (LLMs) to understand proprietary ICPs. ICProFuzz utilizes LLMs to deeply analyze captured communication messages and available partial protocol descriptions. This enables the automated extraction of field semantics, length constraints, and inter-field dependencies. Utilizing the inferred field types and syntactic formats, we design a suite of advanced mutation strategies, including boundary value injection, format-preserving assignment, and protocol-aware keyword insertion. Finally, the generated test cases are fed into the fuzzing engine, where both the semantic parsing and mutation strategies are dynamically adjusted based on test coverage and anomaly feedback. We conducted extensive evaluations using three widely used proprietary ICPs: S7comm, UMAS, and CIP. The experimental results show that ICProFuzz outperforms traditional baseline fuzzers in vulnerability identification, path coverage, and branch coverage. Moreover, ICProFuzz identified 14 vulnerabilities across six real devices, including two zero-day vulnerabilities.

Language models pre-trained on code, such as CodeBERT and CodeGPT, have shown promising performance in software vulnerability detection. However, these models are vulnerable to evasion attacks, in which adversaries apply transformations such as identifier renaming or dead code insertion while maintaining program behaviors, in order to mislead predictions. Although prior research has primarily focused on improving model architectures and code representations to enhance accuracy, the robustness of these systems against evasion attacks is not well studied. In this paper, we propose VulLens, a defense framework compatible with various pre-trained language models to enhance the robustness of software vulnerability detection systems. VulLens begins with code cleaning to remove potentially misleading signals (i.e., malicious dead code), followed by a two-stage detection architecture. In the first stage, high-impact identifiers are selected through attention-based ranking and selectively masked. We then measure the change in prediction confidence between the original and masked inputs: a small change indicates a benign sample, a large change suggests an adversarial sample, and uncertain cases are forwarded to the second stage for further analysis. The second stage employs a self-ensemble strategy enhanced with adaptive diversity-prompting regularization, which explicitly encourages output diversity across multiple ensemble members. By promoting varied decision patterns among the models, this self-ensemble design improves robustness while maintaining flexibility in system deployment and maintenance. Finally, we evaluate VulLens on three pre-trained models across code datasets in Java and C/C++. The results show that VulLens can significantly reduce the attack success rate from 0.867 to 0.223, with minimal impact on detection performance under benign conditions.

Binary vulnerability analysis is increasingly performed by LLM-based agents in an iterative, multi-pass manner, with the model as the core decision-maker. However, how such systems organize exploration over hundreds of reasoning steps remains poorly understood, due to limited context windows and implicit token-level behaviors. We present the first large-scale, trace-level study showing that multi-pass LLM reasoning gives rise to structured, token-level implicit patterns. Analyzing 521 binaries with 99,563 reasoning steps, we identify four dominant patterns: early pruning, path-dependent lock-in, targeted backtracking, and knowledge-guided prioritization that emerge implicitly from reasoning traces. These token-level implicit patterns serve as an abstraction of LLM reasoning: instead of explicit control-flow or predefined heuristics, exploration is organized through implicit decisions regulating path selection, commitment, and revision. Our analysis shows these patterns form a stable, structured system with distinct temporal roles and measurable characteristics. Our results provide the first systematic characterization of LLM-driven binary analysis and a foundation for more reliable analysis systems.

Speech synthesis models can generate highly realistic cloned voices by capturing the phonetic features of a speaker. However, adversaries may exploit this capability by collecting users’ speech samples (e.g., from social media platforms) without their consent to conduct voice cloning or deepfake attacks. While audio adversarial examples have been proposed as a defense mechanism, most of them rely on time-intensive optimization from scratch in the input space, rendering them impractical for latency-sensitive applications such as live audio streaming. They are also vulnerable to countermeasures such as denoising. To address these weaknesses, we propose SpeechShield, a universal and robust perturbation-generation framework that can work offline or in real time. SpeechShield injects adversarial perturbations into speech features that are critical to timbre, such as harmonic structure and energy spectral patterns. It prevents speech synthesis models from learning accurate voice representations and elevates resistance to adversarial removal. Moreover, SpeechShield projects input audio into a compact high-dimensional latent space to generate meta-perturbations, allowing for efficient fine-tuning adaptation to new samples with minimal delay. By incorporating perceptual objectives, SpeechShield also maintains the intelligibility and naturalness of protected speech. We have conducted extensive experiments on a range of advanced models, datasets, and real-world scenarios, showing that SpeechShield consistently outperforms state-of-the-art (SOTA) defenses in terms of protection effectiveness, transferability, and robustness, all at the lowest latency.

Boneh and Franklin introduced one of the earliest constructions of the elegant concept of identity-based encryption (IBE) about two decades ago. Despite significant research advancements and numerous potential applications, IBE has yet to see widespread real-world adoption. A primary challenge is the excessive trust placed in the key derivation server, which inherently possesses the ability to decrypt all ciphertexts. Recent work proposed approaches to reduce server trust in IBE schemes by using a server, called a key curator, that only helps update the public parameters and holds no secret key. In this paper, we present, in the key-curator model, the first IBE scheme that, via a black-box reduction, preserves sender efficiency of a class of IBE schemes (for which we show that there are already well-known and efficient instantiations in the literature), works in the public-random string model (and in particular not requiring extra secret information kept by the key curator) and satisfies ‘sender-efficiency’ performance requirements demanded by many applications (e.g., Internet of Things). All previously proposed IBE schemes with reduced server trust either did not preserve IBE functionality or were subject to server-collusion attacks or only focused on ‘receiver-efficiency’ performance requirements.

Honeypots act as a powerful security mechanism that diverts malicious actors from production systems while providing valuable insights into adversarial behaviors. Yet, the absence of a high-quality honeypot dataset has long impeded robust benchmarking and restricted the employment of advanced AI-driven honeypot defenses. In this work, we address these limitations by constructing a comprehensive shell request-response dataset from Cowrie honeypots. Such a dataset not only facilitates thorough, in-depth honeypot evaluations but also furnishes an essential research foundation for AI-based honeypot development. We analyzed tens of thousands of shell sessions collected during two distinct time frames. (2021-2022 and 2024). By systematically examining command-level usage, session behaviors, and tactics under the MITRE ATT&CK framework, we identified major shifts in adversary techniques, including mounting command complexity, shorter but more targeted infiltration sessions, a more balanced and diverse range of attack methods, and an intensified focus on circumventing defensive measures. These observations emphasize the evolving nature of shell-based intrusions and underscore the necessity for ongoing experimentation and iterative improvements in honeypot design. Through the collection and analysis of this dataset, our work highlights emerging threats in shell defense while also establishing a robust data foundation for the future development of AI-driven honeypots.

Sixth-generation (6G) networks are envisioned to support interconnected local subnetworks that can share specialized, beyond-connectivity services. However, a standardized architecture for discovering and selecting these services across network boundaries has not existed yet. To address this gap, this paper introduces the Central Repository and Selection Function (CRSF), a novel network function for the 6G core that facilitates efficient inter-subnetwork service discovery and selection. We formulate the selection process as a QoS-aware optimization problem designed to balance service quality metrics with user-defined priorities. We evaluate our system model through simulations for a sensing service scenario and observe a consistently higher aggregate Quality of Service (QoS) compared to the baseline selection strategy. The proposed CRSF provides a foundational and extensible mechanism for building standardized, collaborative, and service-centric interconnected networks essential for the 6G era.

Real-time Clock (RTC) has been widely used in various realtime systems to provide precise system time. In this paper, we reveal a new security vulnerability of the RTC circuit, where the internal storage time or timestamp can be arbitrarily modified forward or backward. The security threat of dynamic modifications of system time caused by this vulnerability is called TimeTravel. Based on acoustic resonance and piezoelectric effects, TimeTravel applies acoustic guide waves to the quartz crystal, thereby adjusting the characteristics of the oscillating signal transmitted into the RTC circuit. By manipulating the parameters of acoustic waves, TimeTravel can accelerate or decelerate the timing speed of system time at an adjustable rate, resulting in the relative drift of the timing, which can pose serious safety threats. To assess the severity of TimeTravel, we examine nine modules and seven commercial devices under the RTC circuit. The experimental results show that TimeTravel can drift system time forward and backward at a chosen speed with a maximum 93% accuracy. Our analysis further shows that TimeTravel can maintain an attack success rate of no less than 77% under environments with typical obstacle items.

Federated Learning (FL) exposes vulnerabilities to targeted poisoning attacks that aim to cause misclassification specifically from the source class to the target class. However, using well-established defense frameworks, the poisoning impact of these attacks can be greatly mitigated. We introduce a generalized pre-training stage approach to Boost Targeted Poisoning Attacks against FL, called BoTPA. Its design rationale is to leverage the model update contributions of all data points, including ones outside of the source and target classes, to construct an Amplifier set, in which we falsify the data labels before the FL training process, as a means to boost attacks. We comprehensively evaluate the effectiveness and compatibility of BoTPA on various targeted poisoning attacks. Under data poisoning attacks, our evaluations reveal that BoTPA can achieve a median Relative Increase in Attack Success Rate (RI-ASR) between 15.3% and 36.9% across all possible source-target class combinations, with varying percentages of malicious clients, compared to its baseline. In the context of model poisoning, BoTPA attains RI-ASRs ranging from 13.3% to 94.7% in the presence of the Krum and Multi-Krum defenses, from 2.6% to 49.2% under the Median defense, and from 2.9% to 63.5% under the Flame defense.

The time-delay attacks pose serious security threats to the industrial control systems (ICSs), where ICS infrastructures (e.g., chemical factories) could suffer severe safety consequences. They could bypass current delay detection methods by avoiding triggering packet timeouts. In this article, we reveal that malicious states caused by the time-delay attacks in ICS scenarios can be detected by analyzing ICS programs. We propose detecting a time-delay attack in ICS scenarios by comparing the difference between malicious and benign states, meeting the real-time and noninterference requirements. Specifically, we utilize symbolic execution to analyze ICS programs to generate the benign states of ICS and leverage the key features of time-delay attacks to create the malicious states of ICS, where the states are transferred through the network for remote control and monitoring. We propose a multimodal neural network whose inputs are the malicious states sampled from the ICS network traffic and the time domain features, and the output is whether such a time-delay attack exists. We implement a prototype system and conduct real-world experiments to evaluate the performance of our detection approach. Our experiments cover 102 vulnerable ICS programs and five types of time-delay attacks. The evaluation results show that our approach can detect ICS time-delay attacks in 0.6 s, with 97.2% precision and 98% recall.

Existing registration, discovery, and authorization mechanisms in the 5G core control plane present scalability and efficiency challenges. As cellular deployments scale to accommodate diverse user demands, the 5G core control plane suffers from increased inter-Virtual Network Function (VNF) communication latency, thus deteriorating the reliability of critical procedures. To address this problem, we propose 5G-STREAM (Service mesh Tailored for Reliable, Efficient, and Authorized Microservices) to optimize control plane traffic in distributed cloud environments by establishing a topology awareness of service chains across cloud hierarchies. Leveraging this awareness, 5G-STREAM dynamically configures communication pathways to reduce discovery and authorization signaling overhead, thus increasing the reliability of inter-VNF communication. We develop a prototype of 5G-STREAM and evaluate its performance. Our evaluation results show that 5G-STREAM significantly reduces the process completion time in core service chains by up to 2x inter VNF-Network Repository Function (NRF) latency per transaction, with more pronounced benefits in larger service chains. Furthermore, we show that the cost required to deploy 5G-STREAM is an additional 0.1 USD/hr on AWS for a VNF handling a sustained rate of 50,000 requests/minute.

Existing dynamic vulnerability patching techniques are not well-suited for embedded devices, especially mission-critical ones such as medical equipment, as they have limited computational power and memory but uninterrupted service requirements. Those devices often lack sufficient idle memory for dynamic patching, and the diverse architectures of embedded systems further complicate the creation of patch triggers that are compatible across various system kernels and hardware platforms. To address these challenges, we propose a hot patching framework called StackPatch that facilitates patch development based on stack frame reconstruction. StackPatch introduces different triggering strategies to update programs stored in memory units. We leverage the exception-handling mechanisms commonly available in embedded processors to enhance StackPatch's adaptability across different processor architectures for control flow redirection. We evaluated StackPatch on embedded devices featuring three major microcontroller (MCU) architectures: ARM, RISC-V, and Xtensa. In the experiments, we used StackPatch to successfully fix 102 publicly disclosed vulnerabilities in real-time operating systems (RTOSes). We applied patching to medical devices, soft programmable logic controllers (PLCs), and network services, with StackPatch consistently completing each vulnerability remediation in less than 260 MCU clock cycles.

Large Language Models (LLMs) and their agent systems have recently demonstrated strong potential in automating code reasoning and vulnerability detection. However, when applied to large-scale firmware, their performance degrades due to the binary nature of firmware, complex dependency structures, and heterogeneous components. To address this challenge, this paper presents FIRMHIVE, a recursive agent hive that enables LLMs to act as autonomous firmware security analysts. FIRMHIVE introduces two key mechanisms: (1) transforming delegation into a per-agent, executable primitive and (2) constructing a runtime Tree of Agents (ToA) for decentralized coordination. We evaluate FIRMHIVE using real-world firmware images obtained from publicly available datasets, covering five representative security analysis tasks. Compared with existing LLM-agent baselines, FIRMHIVE performs deeper (about 16x more reasoning steps) and broader (about 2.3x more files inspected) cross-file exploration, resulting in about 5.6x more alerts per firmware. Compared to state-of-the-art (SOTA) security tools, FIRMHIVE identifies about 1.5x more vulnerabilities (1,802 total) and achieves 71

Open Radio Access Network (O-RAN) architectures enhance flexibility for 6G and NextG networks. However, it also brings significant challenges in O-RAN testing with evaluating abundant, high-dimensional key performance indicators (KPIs). In this paper, we introduce a novel two-stage framework to learn temporally-aware low-dimensional representations of O-RAN testing KPIs. To be specific, stage one employs an information-theoretic H-score to train a hybrid self-attentive transformer and echo state network (ESN) reservoir, called Transformer-ESN, capturing temporal dynamics and producing task-aligned 8-dimensional embeddings. Stage two evaluates these embeddings by training a lightweight multilayer perceptron (MLP) predictor exclusively on them for key target KPIs such as reference signal received quality (RSRQ) and spectral efficiency. Using real-world O-RAN testbed data (video streaming with interference), our approach demonstrates a significant advantage specifically when training samples are very limited. In this scenario, the low-dimensional representations learned from the Transformer-ESN yield mean square error (MSE) reductions of up to 41.9% for RSRQ and 29.9% for spectral efficiency compared to predictions from the original high-dimensional data. The framework exhibits high efficiency for O-RAN testing, significantly reducing testing complexities for O-RAN systems.

Actuators are essential components in cyber-physical systems, enabling system modules to perform diverse and complex tasks. Unfortunately, the pursuit of higher functional complexity often correlates with a broader attack surface in actuators. Thus, an efficient automated attack surface assessment is crucial to avoid cyber incidents in critical infrastructures. Limited by enormous parameter spaces, current methods rely on heuristic tests to evaluate interference potential but cannot thoroughly investigate the full spectrum of potential hidden interference. The observation that similar interference trigger configurations lead to the same impact has motivated us to use machine learning algorithms for understanding different impact samples around decision boundaries. By leveraging generalized knowledge of responses against specific attack scenarios, we aim to improve the efficiency of automated attack surface assessment of electromagnetic interference on new targets. To this end, we introduce EMFuzz, an automated mechanism to fuzz hardware to quantify varying adverse effects. We evaluate EMFuzz on 16 new servos within real-world scenarios, where it achieves an 86% accuracy in classifying different attack vectors. With the same test time, EMFuzz uncovers over twice the effective attack configurations of the baseline, greatly improving assessment efficiency. To further validate its efficacy, we apply EMFuzz to assess the attack surface of a new actuator from a robot transfer unit, and it can successfully reveal three distinct adverse effects.

The Open Radio Access Network (O-RAN) architecture enables a flexible, vendor-neutral deployment of 5G networks by disaggregating base station components and supporting third-party xApps for near real-time RAN control. However, the concurrent operation of multiple xApps can lead to conflicting control actions, which may cause network performance degradation. In this work, we propose a framework for xApp conflict management that combines explainable machine learning and causal inference to evaluate the causal relationships between RAN Control Parameters (RCPs) and Key Performance Indicators (KPIs). We use model explainability tools such as SHAP to identify RCPs that jointly affect the same KPI, signaling potential conflicts, and represent these interactions as a causal Directed Acyclic Graph (DAG). We then estimate the causal impact of each of these RCPs on their associated KPIs using metrics such as Average Treatment Effect (ATE) and Conditional Average Treatment Effect (CATE). This approach offers network operators guided insights into identifying conflicts and quantifying their impacts, enabling more informed and effective conflict resolution strategies across diverse xApp deployments.

Machine learning (ML) has been widely used to analyze API call sequences in malware analysis, which typically requires the expertise of domain specialists to extract relevant features from raw data. The extracted features play a critical role in malware analysis. Traditional feature extraction is based on human domain knowledge, while there is a trend of using natural language processing (NLP) for automatic feature extraction. This raises a question: how do we effectively select features for malware analysis based on API call sequences? To answer it, this paper presents a comprehensive study of investigating the impact of feature engineering upon malware classification.We first conducted a comparative performance evaluation under three models, Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and Transformer, with respect to knowledge-based and NLP-based feature engineering methods. We observed that models with knowledge-based feature engineering inputs generally outperform those using NLP-based across all metrics, especially under smaller sample sizes. Then we analyzed a complete set of data features from API call sequences, our analysis reveals that models often focus on features such as handles and virtual addresses, which vary across executions and are difficult for human analysts to interpret.

Fingerprinting is a network reconnaissance technique utilized for gathering information about online computing systems, including operation systems and applications. Unfortunately, attackers typically leverage fingerprinting techniques to locate, enumerate, and subsequently target vulnerable systems, which is the first primary stage of a cyber attack. In this work, we explore the susceptibility of machine learning (ML)-based classifiers to misclassification, where a slight perturbation in the packet is included to spoof OS fingerprints. We propose SOFI (Spoof OS Fingerprints), an adversarial example generation algorithm under TCP/IP specification constraints, to create effective perturbations in a packet for deceiving an OS fingerprint. Specifically, SOFI has three major technical innovations: (1) it is the first to utilize adversarial examples to automatically perturb fingerprinting techniques; (2) it complies with constraints and integrity of network packets; (3) it achieves a high success rate in spoofing OS fingerprints. We validate the effectiveness of adversarial packets against active and passive OS fingerprints, verifying the transferability and robustness of SOFI. Comprehensive experimental results demonstrate that SOFI automatically identifies applicable and available OS fingerprint features, unlike existing tools relying on expert knowledge.