Papers

Data-driven inertial sequence learning has revolutionized navigation in GPS-denied environments, offering superior odometric resolution compared to traditional Bayesian methods. However, deep learning-based inertial tracking systems remain vulnerable to privacy breaches that can expose sensitive training data. \hl{Existing differential privacy solutions often compromise model performance by introducing excessive noise, particularly in high-frequency inertial measurements.} In this article, we propose ConvXformer, a hybrid architecture that fuses ConvNeXt blocks with Transformer encoders in a hierarchical structure for robust inertial navigation. We propose an efficient differential privacy mechanism incorporating adaptive gradient clipping and gradient-aligned noise injection (GANI) to protect sensitive information while ensuring model performance. Our framework leverages truncated singular value decomposition for gradient processing, enabling precise control over the privacy-utility trade-off. Comprehensive performance evaluations on benchmark datasets (OxIOD, RIDI, RoNIN) demonstrate that ConvXformer surpasses state-of-the-art methods, achieving more than 40% improvement in positioning accuracy while ensuring $(\epsilon,\delta)$-differential privacy guarantees. To validate real-world performance, we introduce the Mech-IO dataset, collected from the mechanical engineering building at KAIST, where intense magnetic fields from industrial equipment induce significant sensor perturbations. This demonstrated robustness under severe environmental distortions makes our framework well-suited for secure and intelligent navigation in cyber-physical systems.

As networks move toward the next-generation 6G, Intent-based Management (IbM) systems are increasingly adopted to simplify and automate network management by translating high-level intents into low-level configurations. Within these systems, agents play a critical role in monitoring current state of the network, gathering data, and enforcing actions across the network to fulfill the intent. However, ensuring secure and fine-grained authorization of agents remains a significant challenge, especially in dynamic and multi-tenant environments. Traditional models such as Role-Based Access Control (RBAC), Attribute-Based Access Control (ABAC) and Relational-Based Access Control (RelBAC) often lack the flexibility to accommodate the evolving context and granularity required by intentbased operations. In this paper, we propose an enhanced authorization framework that integrates contextual and functional attributes with agent roles to achieve dynamic, policy-driven access control. By analyzing agent functionalities, our approach ensures that agents are granted only the minimal necessary privileges towards knowledge graphs.

Problem Space: AI Vulnerabilities and Quantum Threats Generative AI vulnerabilities: model inversion, data poisoning, adversarial inputs. Quantum threats Shor Algorithm breaking RSA ECC encryption. Challenge Secure generative AI models against classical and quantum cyberattacks. Proposed Solution Collaborative Penetration Testing Suite Five Integrated Components: DAST SAST OWASP ZAP, Burp Suite, SonarQube, Fortify. IAST Contrast Assess integrated with CI CD pipeline. Blockchain Logging Hyperledger Fabric for tamper-proof logs. Quantum Cryptography Lattice based RLWE protocols. AI Red Team Simulations Adversarial ML & Quantum-assisted attacks. Integration Layer: Unified workflow for AI, cybersecurity, and quantum experts. Key Results 300+ vulnerabilities identified across test environments. 70% reduction in high-severity issues within 2 weeks. 90% resolution efficiency for blockchain-logged vulnerabilities. Quantum-resistant cryptography maintained 100% integrity in tests. Outcome: Quantum AI Security Protocol integrating Blockchain Quantum Cryptography AI Red Teaming.

Wireless Body Area Networks (WBANs) have gained significant attention due to their applications in healthcare monitoring, sports, military communication, and remote patient care. These networks consist of wearable or implanted sensors that continuously collect and transmit physiological data, requiring efficient and reliable communication. However, WBANs face challenges such as limited energy, dynamic topology, and sensitivity to node temperature, which demand specialized routing strategies. Traditional shortest-path routing often causes congestion and overheating in specific nodes, leading to early failures. To address these problems, this paper proposes an intelligent temperature-aware and reliability-based routing approach that enhances WBAN performance. The proposed method works in two phases: (1) network setup and intelligent path selection, and (2) dynamic traffic management and hotspot avoidance. In the first phase, nodes share information such as residual energy, temperature, link reliability, and delay to build an optimized topology using a multi-criteria decision algorithm. The second phase continuously monitors real-time conditions and reroutes traffic away from overheated or depleted nodes. Simulation results show that the proposed approach improves throughput by 13 percent, reduces end-to-end delay by 10 percent, decreases energy consumption by 25 percent, and lowers routing load by 30 percent compared to existing methods.

The increasing reliance on AI-driven 5G/6G network infrastructures for mission-critical services highlights the need for reliability and resilience against sophisticated cyber-physical threats. These networks are highly exposed to novel attack surfaces due to their distributed intelligence, virtualized resources, and cross-domain integration. This paper proposes a fault-tolerant and resilience-aware framework that integrates AI-driven anomaly detection, adaptive routing, and redundancy mechanisms to mitigate cascading failures under cyber-physical attack conditions. A comprehensive validation is carried out using NS-3 simulations, where key performance indicators such as reliability, latency, resilience index, and packet loss rate are analyzed under various attack scenarios. The deduced results demonstrate that the proposed framework significantly improves fault recovery, stabilizes packet delivery, and reduces service disruption compared to baseline approaches.

Objectives: This study aims to investigate the readability and understandability of bitwise operators in programming, with the main hypothesis that there will be a difference in the performance metrics (response time and error rate) between participants exposed to various bitwise operators related questions and those who are not. Participants: Participants in this human research study include people without programming background, novice programmers, and university students with varying programming experience (from freshmen to PhD level). There were 23 participants in this study. Study Methods: This study uses a within-subjects experimental design to assess how people with diverse programming backgrounds understand and use bitwise operators. Participants complete tasks in a JavaScript program, and their task completion times and task accuracy are recorded for analysis. Findings: The results indicate that operators can be one of the factors predicting response time, showing a small but significant effect (R-squared = 0.032, F(1, 494) = 16.5, p < .001). Additionally, operators such as OR, NOT, and Left Shift showed statistical significance in task completion times compared to other operators. Conclusions: While the complexity of bitwise operators did not generally result in longer task completion times, certain operators were found to be less intuitive, suggesting the need for further investigation and potential redesign for improved understandability.

With the goal of improving the security of Internet protocols, we seek faster, semi-automatic methods to discover new vulnerabilities in protocols such as DNS, BGP, and others. To this end, we introduce the LLM-Assisted Protocol Attack Discovery (LAPRAD) methodology, enabling security researchers with some DNS knowledge to efficiently uncover vulnerabilities that would otherwise be hard to detect. LAPRAD follows a three-stage process. In the first, we consult an LLM (GPT-o1) that has been trained on a broad corpus of DNS-related sources and previous DDoS attacks to identify potential exploits. In the second stage, a different LLM automatically constructs the corresponding attack configurations using the ReACT approach implemented via LangChain (DNS zone file generation). Finally, in the third stage, we validate the attack's functionality and effectiveness. Using LAPRAD, we uncovered three new DDoS attacks on the DNS protocol and rediscovered two recently reported ones that were not included in the LLM's training data. The first new attack employs a bait-and-switch technique to trick resolvers into caching large, bogus DNSSEC RRSIGs, reducing their serving capacity to as little as 6%. The second exploits large DNSSEC encryption algorithms (RSA-4096) with multiple keys, thereby bypassing a recently implemented default RRSet limit. The third leverages ANY-type responses to produce a similar effect. These variations of a cache-flushing DDoS attack, called SigCacheFlush, circumvent existing patches, severely degrade resolver query capacity, and impact the latest versions of major DNS resolver implementations.

When large language model (LLM) agents are increasingly deployed to automate tasks and interact with untrusted external data, prompt injection emerges as a significant security threat. By injecting malicious instructions into the data that LLMs access, an attacker can arbitrarily override the original user task and redirect the agent toward unintended, potentially harmful actions. Existing defenses either require access to model weights (fine-tuning), incur substantial utility loss (detection-based), or demand non-trivial system redesign (system-level). Motivated by this, we propose DataFilter, a test-time model-agnostic defense that removes malicious instructions from the data before it reaches the backend LLM. DataFilter is trained with supervised fine-tuning on simulated injections and leverages both the user's instruction and the data to selectively strip adversarial content while preserving benign information. Across multiple benchmarks, DataFilter consistently reduces the prompt injection attack success rates to near zero while maintaining the LLMs' utility. DataFilter delivers strong security, high utility, and plug-and-play deployment, making it a strong practical defense to secure black-box commercial LLMs against prompt injection. Our DataFilter model is released at https://huggingface.co/JoyYizhu/DataFilter for immediate use, with the code to reproduce our results at https://github.com/yizhu-joy/DataFilter.

As large language models (LLMs) are increasingly integrated into real-world applications, ensuring their safety, robustness, and privacy compliance has become critical. We present OpenGuardrails, the first fully open-source platform that unifies large-model-based safety detection, manipulation defense, and deployable guardrail infrastructure. OpenGuardrails protects against three major classes of risks: (1) content-safety violations such as harmful or explicit text generation, (2) model-manipulation attacks including prompt injection, jailbreaks, and code-interpreter abuse, and (3) data leakage involving sensitive or private information. Unlike prior modular or rule-based frameworks, OpenGuardrails introduces three core innovations: (1) a Configurable Policy Adaptation mechanism that allows per-request customization of unsafe categories and sensitivity thresholds; (2) a Unified LLM-based Guard Architecture that performs both content-safety and manipulation detection within a single model; and (3) a Quantized, Scalable Model Design that compresses a 14B dense base model to 3.3B via GPTQ while preserving over 98 of benchmark accuracy. The system supports 119 languages, achieves state-of-the-art performance across multilingual safety benchmarks, and can be deployed as a secure gateway or API-based service for enterprise use. All models, datasets, and deployment scripts are released under the Apache 2.0 license.

The growing use of third-party hardware accelerators (e.g., FPGAs, ASICs) for deep neural networks (DNNs) introduces new security vulnerabilities. Conventional model-level backdoor attacks, which only poison a model's weights to misclassify inputs with a specific trigger, are often detectable because the entire attack logic is embedded within the model (i.e., software), creating a traceable layer-by-layer activation path. This paper introduces the HArdware-Model Logically Combined Attack (HAMLOCK), a far stealthier threat that distributes the attack logic across the hardware-software boundary. The software (model) is now only minimally altered by tuning the activations of few neurons to produce uniquely high activation values when a trigger is present. A malicious hardware Trojan detects those unique activations by monitoring the corresponding neurons' most significant bit or the 8-bit exponents and triggers another hardware Trojan to directly manipulate the final output logits for misclassification. This decoupled design is highly stealthy, as the model itself contains no complete backdoor activation path as in conventional attacks and hence, appears fully benign. Empirically, across benchmarks like MNIST, CIFAR10, GTSRB, and ImageNet, HAMLOCK achieves a near-perfect attack success rate with a negligible clean accuracy drop. More importantly, HAMLOCK circumvents the state-of-the-art model-level defenses without any adaptive optimization. The hardware Trojan is also undetectable, incurring area and power overheads as low as 0.01%, which is easily masked by process and environmental noise. Our findings expose a critical vulnerability at the hardware-software interface, demanding new cross-layer defenses against this emerging threat.

This report presents a comprehensive analysis of an unsupervised multi-expert machine learning framework for detecting short ticketing fraud in railway systems. The study introduces an A/B/C/D station classification system that successfully identifies suspicious patterns across 30 high-risk stations. The framework employs four complementary algorithms: Isolation Forest, Local Outlier Factor, One-Class SVM, and Mahalanobis Distance. Key findings include the identification of five distinct short ticketing patterns and potential for short ticketing recovery in transportation systems.

Cyberattack detection in Critical Infrastructure and Supply Chains has become challenging in Industry 4.0. Intrusion Detection Systems (IDS) are deployed to counter the cyberattacks. However, an IDS effectively detects attacks based on the known signatures and patterns, Zero-day attacks go undetected. To overcome this drawback in IDS, the integration of a Dense Neural Network (DNN) with Data Augmentation is proposed. It makes IDS intelligent and enables it to self-learn with high accuracy when a novel attack is encountered. The network flow captures datasets are highly imbalanced same as the real network itself. The Data Augmentation plays a crucial role in balancing the data. The balancing of data is challenging as the minority class is as low as 0.000004% of the dataset, and the abundant class is higher than 80% of the dataset. Synthetic Minority Oversampling Technique is used for balancing the data. However, higher accuracies are achieved with balanced test data, lower accuracies are noticeable with the original imbalanced test data suggesting overfitting. A comparison with state-of-the-art research using Synthetic Minority Oversampling Technique with Edited Nearest Neighbor shows the classification of classes remains poor for the original dataset. This suggests highly imbalanced datasets of network flow require a different method of data augmentation.

In recent years, blockchain has experienced widespread adoption across various industries, becoming integral to numerous enterprise applications. Concurrently, the rise of generative AI and LLMs has transformed human-computer interactions, offering advanced capabilities in understanding and generating human-like text. The introduction of the MCP has further enhanced AI integration by standardizing communication between AI systems and external data sources. Despite these advancements, there is still no standardized method for seamlessly integrating LLM applications and blockchain. To address this concern, we propose "MCC: Model Context Contracts" a novel framework that enables LLMs to interact directly with blockchain smart contracts through MCP-like protocol. This integration allows AI agents to invoke blockchain smart contracts, facilitating more dynamic and context-aware interactions between users and blockchain networks. Essentially, it empowers users to interact with blockchain systems and perform transactions using queries in natural language. Within this proposed architecture, blockchain smart contracts can function as intelligent agents capable of recognizing user input in natural language and executing the corresponding transactions. To ensure that the LLM accurately interprets natural language inputs and maps them to the appropriate MCP functions, the LLM was fine-tuned using a custom dataset comprising user inputs paired with their corresponding MCP server functions. This fine-tuning process significantly improved the platform's performance and accuracy. To validate the effectiveness of MCC, we have developed an end-to-end prototype implemented on the Rahasak blockchain with the fine-tuned Llama-4 LLM. To the best of our knowledge, this research represents the first approach to using the concept of Model Context Protocol to integrate LLMs with blockchain.

Recent findings suggest that misaligned models may exhibit deceptive behavior, raising concerns about output trustworthiness. Chain-of-thought (CoT) is a promising tool for alignment monitoring: when models articulate their reasoning faithfully, monitors can detect and mitigate harmful behaviors before undesirable outcomes occur. However, a key uncertainty is: Can models obfuscate their CoT in order to pursue hidden adversarial objectives while evading detection? To answer this question and thus stress-test CoT monitorability, we develop a composable and quantifiable taxonomy of prompts to elicit CoT obfuscation. We evaluate both internal CoT (reasoning traces) and external CoT (prompted reasoning in outputs) using toy tasks and more realistic environments in SHADE-Arena. We show that: (i) CoT monitoring performs accurately and efficiently without obfuscation pressure. (ii) Under strong obfuscation pressure, some models successfully complete adversarial tasks while evading detection. (iii) Models do not obfuscate their internal CoT as much as their external CoT (under prompt pressure). These results suggest that while CoT provides valuable oversight in benign settings, robust deployment requires model-specific stress-testing of monitorability.

The rapid growth of the Internet of Things (IoT) has transformed industries by enabling seamless data exchange among connected devices. However, IoT networks remain vulnerable to security threats such as denial of service (DoS) attacks, anomalous traffic, and data manipulation due to decentralized architectures and limited resources. To address these issues, this paper proposes an advanced anomaly detection framework with three main phases. First, data preprocessing is performed using the Median KS Test to remove noise, handle missing values, and balance datasets for cleaner input. Second, a feature selection phase employs a Genetic Algorithm combined with eagle inspired search strategies to identify the most relevant features, reduce dimensionality, and improve efficiency without sacrificing accuracy. Finally, an ensemble classifier integrates Decision Tree, Random Forest, and XGBoost algorithms to achieve accurate and reliable anomaly detection. The proposed model demonstrates high adaptability and scalability across diverse IoT environments. Experimental results show that it outperforms existing methods by achieving 98 percent accuracy, 95 percent detection rate, and reductions in false positive (10 percent) and false negative (5 percent) rates. These results confirm the framework effectiveness and robustness in improving IoT network security against evolving cyber threats.

We show that if a language $L$ admits a public-coin unambiguous interactive proof (UIP) with round complexity $\ell$, where $a$ bits are communicated per round, then the batch language $L^{\otimes k}$, i.e. the set of $k$-tuples of statements all belonging to $L$, has an unambiguous interactive proof with round complexity $\ell\cdot\mathsf{polylog}(k)$, per-round communication of $a\cdot \ell\cdot\mathsf{polylog}(k) + \mathsf{poly}(\ell)$ bits, assuming the verifier in the $\mathsf{UIP}$ has depth bounded by $\mathsf{polylog}(k)$. Prior to this work, the best known batch $\mathsf{UIP}$ for $L^{\otimes{k}}$ required communication complexity at least $(\mathsf{poly}(a)\cdot k^{\epsilon} + k) \cdot \ell^{1/\epsilon}$ for any arbitrarily small constant $\epsilon>0$ (Reingold-Rothblum-Rothblum, STOC 2016). As a corollary of our result, we obtain a doubly efficient proof system, that is, a proof system whose proving overhead is polynomial in the time of the underlying computation, for any language computable in polynomial space and in time at most $n^{O\left(\sqrt{\frac{\log n}{\log\log n}}\right)}$. This expands the state of the art of doubly efficient proof systems: prior to our work, such systems were known for languages computable in polynomial space and in time $n^{({\log n})^\delta}$ for a small $\delta>0$ significantly smaller than $1/2$ (Reingold-Rothblum-Rothblum, STOC 2016).

In recent years, the rapid integration of Internet of Things (IoT) devices into the healthcare sector has brought about revolutionary advancements in patient care and data management. While these technological innovations hold immense promise, they concurrently raise critical security concerns, particularly in safeguarding medical data against potential cyber threats. The sensitive nature of health-related information requires robust measures to ensure the confidentiality, integrity, and availability of patient data in IoT-enabled medical environments. Addressing the imperative need for enhanced security in IoT-based healthcare systems, we propose a comprehensive method encompassing three distinct phases. In the first phase, we implement Blockchain-Enabled Request and Transaction Encryption to strengthen data transaction security, providing an immutable and transparent framework. In the second phase, we introduce a Request Pattern Recognition Check that leverages diverse data sources to identify and block potential unauthorized access attempts. Finally, the third phase incorporates Feature Selection and a BiLSTM network to enhance the accuracy and efficiency of intrusion detection using advanced machine learning techniques. We compared the simulation results of the proposed method with three recent related methods: AIBPSF-IoMT, OMLIDS-PBIoT, and AIMMFIDS. The evaluation criteria include detection rate, false alarm rate, precision, recall, and accuracy - crucial benchmarks for assessing the overall performance of intrusion detection systems. Our findings show that the proposed method outperforms existing approaches across all evaluated criteria, demonstrating its effectiveness in improving the security of IoT-based healthcare systems.

We investigate and defend the possibility of causing a stock market crash via small manipulations of individual stock values that together realize an adversarial example to financial forecasting models, causing these models to make the self-fulfilling prediction of a crash. Such a crash triggered by an adversarial example would likely be hard to detect, since the model's predictions would be accurate and the interventions that would cause it are minor. This possibility is a major risk to financial stability and an opportunity for hostile actors to cause great economic damage to an adversary. This threat also exists against individual stocks and the corresponding valuation of individual companies. We outline how such an attack might proceed, what its theoretical basis is, how it can be directed towards a whole economy or an individual company, and how one might defend against it. We conclude that this threat is vastly underappreciated and requires urgent research on how to defend against it.

Disaggregated storage with NVMe-over-Fabrics (NVMe-oF) has emerged as the standard solution in modern supercomputers and data center clusters, achieving superior performance, resource utilization, and power efficiency. Simultaneously, confidential computing (CC) is becoming the de facto security paradigm, enforcing stronger isolation and protection for sensitive workloads. However, securing state-of-the-art storage with traditional CC methods struggles to scale and compromises performance or security. To address these issues, we introduce Hazel, a storage management system that extends the NVMe-oF protocol capabilities and adheres to the CC threat model, providing confidentiality, integrity, and freshness guarantees. Hazel offers an appropriate control path with novel concepts such as counter-leasing. Hazel also optimizes data path performance by leveraging NVMe metadata and introducing a new disaggregated Hazel Merkle Tree (HMT), all while remaining compatible with NVMe-oF. For additional efficiency, Hazel also supports offloading to CC-capable smart NIC accelerators. We prototype Hazel on an NVIDIA BlueField-3 and demonstrate that it can achieve as little as 1-2% performance degradation for synthetic patterns, AI training, IO500, and YCSB.

Large Language Models (LLMs) remain vulnerable to multi-turn jailbreak attacks. We introduce HarmNet, a modular framework comprising ThoughtNet, a hierarchical semantic network; a feedback-driven Simulator for iterative query refinement; and a Network Traverser for real-time adaptive attack execution. HarmNet systematically explores and refines the adversarial space to uncover stealthy, high-success attack paths. Experiments across closed-source and open-source LLMs show that HarmNet outperforms state-of-the-art methods, achieving higher attack success rates. For example, on Mistral-7B, HarmNet achieves a 99.4% attack success rate, 13.9% higher than the best baseline. Index terms: jailbreak attacks; large language models; adversarial framework; query refinement.

International students (IntlS) in the US refer to foreign students who acquire student visas to study in the US, primarily in higher education. As IntlS arrive in the US, they face several challenges, such as adjusting to a new country and culture, securing housing remotely, and arranging finances for tuition and personal expenses. These experiences, coupled with recent events such as visa revocations and the cessation of new visas, compound IntlS' risk of being targeted by and falling victim to online scams. While prior work has investigated IntlS' security and privacy, as well as general end users' reactions to online scams, research on how IntlS are uniquely impacted by scams remains largely absent. To address this gap, we conduct a two-phase user study comprising surveys (n=48) and semi-structured interviews (n=9). We investigate IntlS' exposure and interactions with scams, post-exposure actions such as reporting, and their perceptions of the usefulness of existing prevention resources and the barriers to following prevention advice. We find that IntlS are often targeted by scams (e.g., attackers impersonating government officials) and fear legal implications or deportation, which directly impacts their interactions with scams (e.g., they may prolong engagement with a scammer due to a sense of urgency). Interestingly, we also find that IntlS may lack awareness of - or access to - reliable resources that inform them about scams or guide them in reporting incidents to authorities. In fact, they may also face unique barriers in enacting scam prevention advice, such as avoiding reporting financial losses, since IntlS are required to demonstrate financial ability to stay in the US. The findings produced by our study help synthesize guidelines for stakeholders to better aid IntlS in reacting to scams.

As large language models (LLMs) become progressively more embedded in clinical decision-support, documentation, and patient-information systems, ensuring their privacy and trustworthiness has emerged as an imperative challenge for the healthcare sector. Fine-tuning LLMs on sensitive electronic health record (EHR) data improves domain alignment but also raises the risk of exposing patient information through model behaviors. In this work-in-progress, we present an exploratory empirical study on membership inference vulnerabilities in clinical LLMs, focusing on whether adversaries can infer if specific patient records were used during model training. Using a state-of-the-art clinical question-answering model, Llemr, we evaluate both canonical loss-based attacks and a domain-motivated paraphrasing-based perturbation strategy that more realistically reflects clinical adversarial conditions. Our preliminary findings reveal limited but measurable membership leakage, suggesting that current clinical LLMs provide partial resistance yet remain susceptible to subtle privacy risks that could undermine trust in clinical AI adoption. These results motivate continued development of context-aware, domain-specific privacy evaluations and defenses such as differential privacy fine-tuning and paraphrase-aware training, to strengthen the security and trustworthiness of healthcare AI systems.

E-values have gained prominence as flexible tools for statistical inference and risk control, enabling anytime- and post-hoc-valid procedures under minimal assumptions. However, many real-world applications fundamentally rely on sensitive data, which can be leaked through e-values. To ensure their safe release, we propose a general framework to transform non-private e-values into differentially private ones. Towards this end, we develop a novel biased multiplicative noise mechanism that ensures our e-values remain statistically valid. We show that our differentially private e-values attain strong statistical power, and are asymptotically as powerful as their non-private counterparts. Experiments across online risk monitoring, private healthcare, and conformal e-prediction demonstrate our approach's effectiveness and illustrate its broad applicability.

Networked Control Systems (NCSs) are integral in critical infrastructures such as power grids, transportation networks, and production systems. Ensuring the resilient operation of these large-scale NCSs against cyber-attacks is crucial for societal well-being. Over the past two decades, extensive research has been focused on developing metrics to quantify the vulnerabilities of NCSs against attacks. Once the vulnerabilities are quantified, mitigation strategies can be employed to enhance system resilience. This article provides a comprehensive overview of methods developed for assessing NCS vulnerabilities and the corresponding mitigation strategies. Furthermore, we emphasize the importance of probabilistic risk metrics to model vulnerabilities under adversaries with imperfect process knowledge. The article concludes by outlining promising directions for future research.

We present Qatsi, a hierarchical key derivation scheme using Argon2id that generates reproducible cryptographic secrets without persistent storage. The system eliminates vault-based attack surfaces by deriving all secrets deterministically from a single high-entropy master secret and contextual layers. Outputs achieve 103-312 bits of entropy through memory-hard derivation (64-128 MiB, 16-32 iterations) and provably uniform rejection sampling over 7776-word mnemonics or 90-character passwords. We formalize the hierarchical construction, prove output uniformity, and quantify GPU attack costs: $2.4 \times 10^{16}$ years for 80-bit master secrets on single-GPU adversaries under Paranoid parameters (128 MiB memory). The implementation in Rust provides automatic memory zeroization, compile-time wordlist integrity verification, and comprehensive test coverage. Reference benchmarks on Apple M1 Pro (2021) demonstrate practical usability with 544 ms Standard mode and 2273 ms Paranoid mode single-layer derivations. Qatsi targets air-gapped systems and master credential generation where stateless reproducibility outweighs rotation flexibility.

RISC-V processors are becoming ubiquitous in critical applications, but their susceptibility to microarchitectural side-channel attacks is a serious concern. Detection of microarchitectural attacks in RISC-V is an emerging research topic that is relatively underexplored, compared to x86 and ARM. The first line of work to detect flush+fault-based microarchitectural attacks in RISC-V leverages Machine Learning (ML) models, yet it leaves several practical aspects that need further investigation. To address overlooked issues, we leveraged gem5 and propose a new detection method combining statistical preprocessing and association rule mining having reconfiguration capabilities to generalize the detection method for any microarchitectural attack. The performance comparison with state-of-the-art reveals that the proposed detection method achieves up to 5.15% increase in accuracy, 7% rise in precision, and 3.91% improvement in recall under the cryptographic, computational, and memory-intensive workloads alongside its flexibility to detect new variant of flush+fault attack. Moreover, as the attack detection relies on association rules, their human-interpretable nature provides deep insight to understand microarchitectural behavior during the execution of attack and benign applications.

Mobile apps often embed authentication secrets, such as API keys, tokens, and client IDs, to integrate with cloud services. However, developers often hardcode these credentials into Android apps, exposing them to extraction through reverse engineering. Once compromised, adversaries can exploit secrets to access sensitive data, manipulate resources, or abuse APIs, resulting in significant security and financial risks. Existing detection approaches, such as regex-based analysis, static analysis, and machine learning, are effective for identifying known patterns but are fundamentally limited: they require prior knowledge of credential structures, API signatures, or training data. In this paper, we propose SecretLoc, an LLM-based approach for detecting hardcoded secrets in Android apps. SecretLoc goes beyond pattern matching; it leverages contextual and structural cues to identify secrets without relying on predefined patterns or labeled training sets. Using a benchmark dataset from the literature, we demonstrate that SecretLoc detects secrets missed by regex-, static-, and ML-based methods, including previously unseen types of secrets. In total, we discovered 4828 secrets that were undetected by existing approaches, discovering more than 10 "new" types of secrets, such as OpenAI API keys, GitHub Access Tokens, RSA private keys, and JWT tokens, and more. We further extend our analysis to newly crawled apps from Google Play, where we uncovered and responsibly disclosed additional hardcoded secrets. Across a set of 5000 apps, we detected secrets in 2124 apps (42.5%), several of which were confirmed and remediated by developers after we contacted them. Our results reveal a dual-use risk: if analysts can uncover these secrets with LLMs, so can attackers. This underscores the urgent need for proactive secret management and stronger mitigation practices across the mobile ecosystem.

Phishing websites remain a significant cybersecurity threat, necessitating accurate and cost-effective detection mechanisms. In this paper, we present CLASP, a novel system that effectively identifies phishing websites by leveraging multiple intelligent agents, built using large language models (LLMs), to analyze different aspects of a web resource. The system processes URLs or QR codes, employing specialized LLM-based agents that evaluate the URL structure, webpage screenshot, and HTML content to predict potential phishing threats. To optimize performance while minimizing operational costs, we experimented with multiple combination strategies for agent-based analysis, ultimately designing a strategic combination that ensures the per-website evaluation expense remains minimal without compromising detection accuracy. We tested various LLMs, including Gemini 1.5 Flash and GPT-4o mini, to build these agents and found that Gemini 1.5 Flash achieved the best performance with an F1 score of 83.01% on a newly curated dataset. Also, the system maintained an average processing time of 2.78 seconds per website and an API cost of around $3.18 per 1,000 websites. Moreover, CLASP surpasses leading previous solutions, achieving over 40% higher recall and a 20% improvement in F1 score for phishing detection on the collected dataset. To support further research, we have made our dataset publicly available, supporting the development of more advanced phishing detection systems.

The DNS infrastructure is infamous for facilitating reflective amplification attacks. Various countermeasures such as server shielding, access control, rate limiting, and protocol restrictions have been implemented. Still, the threat remains throughout the deployment of DNS servers. In this paper, we report on and evaluate the often unnoticed threat that derives from transparent DNS forwarders, a widely deployed, incompletely functional set of DNS components. Transparent DNS forwarders transfer DNS requests without rebuilding packets with correct source addresses. As such, transparent forwarders feed DNS requests into (mainly powerful and anycasted) open recursive resolvers, which thereby can be misused to participate unwillingly in distributed reflective amplification attacks. We show how transparent forwarders raise severe threats to the Internet infrastructure. They easily circumvent rate limiting and achieve an additional, scalable impact via the DNS anycast infrastructure. We empirically verify this scaling behavior up to a factor of 14. Transparent forwarders can also assist in bypassing firewall rules that protect recursive resolvers, making these shielded infrastructure entities part of the global DNS attack surface.

The integration of Internet of Things (IoT) devices in healthcare has revolutionized patient care by enabling real-time monitoring, personalized treatments, and efficient data management. However, this technological advancement introduces significant security risks, particularly concerning the confidentiality, integrity, and availability of sensitive medical data. Traditional security measures are often insufficient to address the unique challenges posed by IoT environments, such as heterogeneity, resource constraints, and the need for real-time processing. To tackle these challenges, we propose a comprehensive three-phase security framework designed to enhance the security and reliability of IoT-enabled healthcare systems. In the first phase, the framework assesses the reliability of IoT devices using a reputation-based trust estimation mechanism, which combines device behavior analytics with off-chain data storage to ensure scalability. The second phase integrates blockchain technology with a lightweight proof-of-work mechanism, ensuring data immutability, secure communication, and resistance to unauthorized access. The third phase employs a lightweight Long Short-Term Memory (LSTM) model for anomaly detection and classification, enabling real-time identification of cyber threats. Simulation results demonstrate that the proposed framework outperforms existing methods, achieving a 2% increase in precision, accuracy, and recall, a 5% higher attack detection rate, and a 3% reduction in false alarm rate. These improvements highlight the framework's ability to address critical security concerns while maintaining scalability and real-time performance.