Papers

Multi-agent systems face a fundamental architectural flaw: agent identity is bound to network location. When agents migrate between providers, scale across instances, or federate across organizations, URI-based identity schemes break references, fragment audit trails, and require centralized coordination. We propose the agent:// URI scheme, which decouples identity from topology through three orthogonal components: a trust root establishing organizational authority, a hierarchical capability path enabling semantic discovery, and a sortable unique identifier providing stable reference. The scheme enables capability-based discovery through DHT key derivation, where queries return agents by what they do rather than where they are. Trust-root scoping prevents cross-organization pollution while permitting federation when desired. Cryptographic attestation via PASETO tokens binds capability claims to agent identity, enabling verification without real-time contact with the issuing authority. We evaluate the scheme across four dimensions: capability expressiveness (100% coverage on 369 production tools with zero collision), discovery precision (F1=1.0 across 10,000 agents), identity stability (formal proofs of migration invariance), and performance (all operations under 5 microseconds). The agent:// URI scheme provides a formally-specified, practically-evaluated foundation for decentralized agent identity and capability-based discovery.

MITRE ATT&CK is a cybersecurity knowledge base that organizes threat actor and cyber-attack information into a set of tactics describing the reasons and goals threat actors have for carrying out attacks, with each tactic having a set of techniques that describe the potential methods used in these attacks. One major application of ATT&CK is the use of its tactic and technique hierarchy by security specialists as a framework for annotating cyber-threat intelligence reports, vulnerability descriptions, threat scenarios, inter alia, to facilitate downstream analyses. To date, the tagging process is still largely done manually. In this technical note, we provide a stratified "task space" characterization of the MITRE ATT&CK text tagging task for organizing previous efforts toward automation using AIML methods, while also clarifying pathways for constructing new methods. To illustrate one of the pathways, we use the task space strata to stage-wise construct our own multi-label hierarchical classification models for the text tagging task via experimentation over general cyber-threat intelligence text – using shareable computational tools and publicly releasing the models to the security community (via https://github.com/jpmorganchase/MITRE_models). Our multi-label hierarchical approach yields accuracy scores of roughly 94% at the tactic level, as well as accuracy scores of roughly 82% at the technique level. The models also meet or surpass state-of-the-art performance while relying only on classical machine learning methods – removing any dependence on LLMs, RAG, agents, or more complex hierarchical approaches. Moreover, we show that GPT-4o model performance at the tactic level is significantly lower (roughly 60% accuracy) than our own approach. We also extend our baseline model to a corpus of threat scenarios for financial applications produced by subject matter experts.

As the expansion of IoT connectivity continues to provide quality-of-life improvements around the world, they simultaneously introduce increasing privacy and security concerns. The lack of a clear definition in managing shared and protected access to IoT sensors offer channels by which devices can be compromised and sensitive data can be leaked. In recent years, WebAssembly has received considerable attention for its efficient application sandboxing suitable for embedded systems, making it a prime candidate for exploring a secure and portable sensor interface. This paper introduces the first WebAssembly System Interface (WASI) extension offering a secure, portable, and low-footprint sandbox enabling multi-tenant access to sensor data across heterogeneous embedded devices. The runtime extensions provide application memory isolation, ensure appropriate resource privileges by intercepting sensor access, and offer an MQTT-SN interface enabling in-network access control. When targeting the WebAssembly byte-code with the associated runtime extensions implemented atop the Zephyr RTOS, our evaluation of sensor access indicates a latency overhead of 6% with an additional memory footprint of 5% when compared to native execution. As MQTT-SN requests are dominated by network delays, the WASI-SN implementation of MQTT-SN introduces less than 1% additional latency with similar memory footprint.

Quantum Key Distribution (QKD) allows secure communication without relying on computational assumptions, but can currently only be deployed over relatively short distances due to hardware constraints. To extend QKD over long distances, networks of trusted repeater nodes can be used, wherein QKD is executed between neighbouring nodes and messages between non-neighbouring nodes are forwarded using a relay protocol. Although these networks are being deployed worldwide, no protocol exists which provides provable guarantees of integrity against manipulation from both external adversaries and corrupted intermediates. In this work, we present the first protocol that provably provides both confidentiality and integrity. Our protocol combines an existing cryptographic technique, Algebraic Manipulation Detection (AMD) codes, with multi-path relaying over trusted repeater networks. This protocol achieves Information Theoretic Security (ITS) against the detection of manipulation, which we prove formally through a sequence of games.

Accurate Android malware detection was critical for protecting users at scale. Signature scanners lagged behind fast release cycles on public app stores. We aimed to build a trustworthy detector by pairing a comprehensive dataset with a rigorous, transparent evaluation, and to identify interpretable drivers of decisions. We used CICMalDroid2020, which contained 17,341 apps across Benign, Adware, Banking, SMS malware, and Riskware. We extracted 301 static and 263 dynamic features into a 564 dimensional hybrid vector, then evaluated seven classifiers under three schemes, original features, principal component analysis, PCA, and linear discriminant analysis, LDA, with a 70 percent training and 30 percent test split. Results showed that gradient boosting on the original features performed best. XGBoost achieved 0.9747 accuracy, 0.9703 precision, 0.9731 recall, and 0.9716 F1, and the confusion matrix indicated rare benign labels for malicious apps. HistGradientBoosting reached 0.9741 accuracy and 0.9708 F1, while CatBoost and Random Forest were slightly lower at 0.9678 and 0.9687 accuracy with 0.9636 and 0.9637 F1. KNN and SVM lagged. PCA reduced performance for all models, with XGBoost dropping to 0.9164 accuracy and 0.8988 F1. LDA maintained mid 90s accuracy and clarified separable clusters in projections. A depth two surrogate tree highlighted package name, main activity, and target SDK as key drivers. These findings established high fidelity supervised baselines for Android malware detection and indicated that rich hybrid features with gradient boosting offered a practical and interpretable foundation for deployment.

The deployment of large language models (LLMs) has raised security concerns due to their susceptibility to producing harmful or policy-violating outputs when exposed to adversarial prompts. While alignment and guardrails mitigate common misuse, they remain vulnerable to automated jailbreaking methods such as GCG, PEZ, and GBDA, which generate adversarial suffixes via training and gradient-based search. Although effective, these methods particularly GCG are computationally expensive, limiting their practicality for organisations with constrained resources. This paper introduces a resource-efficient adversarial prompting approach that eliminates the need for retraining by matching new prompts to a database of pre-trained adversarial prompts. A dataset of 1,000 prompts was classified into seven harm-related categories, and GCG, PEZ, and GBDA were evaluated on a Llama 3 8B model to identify the most effective attack method per category. Results reveal a correlation between prompt type and algorithm effectiveness. By retrieving semantically similar successful adversarial prompts, the proposed method achieves competitive attack success rates with significantly reduced computational cost. This work provides a practical framework for scalable red-teaming and security evaluation of aligned LLMs, including in settings where model internals are inaccessible.

In an era where cyber threats are rapidly evolving, the reliability of cyber forensic analysis has become increasingly critical for effective digital investigations and cybersecurity responses. AI agents are being adopted across digital forensic practices due to their ability to automate processes such as anomaly detection, evidence classification, and behavioral pattern recognition, significantly enhancing scalability and reducing investigation timelines. However, the characteristics that make AI indispensable also introduce notable risks. AI systems, often trained on biased or incomplete datasets, can produce misleading results, including false positives and false negatives, thereby jeopardizing the integrity of forensic investigations. This study presents a meticulous comparative analysis of the effectiveness of the most used AI agent, ChatGPT, and human forensic investigators in the realm of cyber forensic analysis. Our research reveals critical limitations within AI-driven approaches, demonstrating scenarios in which sophisticated or novel cyber threats remain undetected due to the rigid pattern-based nature of AI systems. Conversely, our analysis highlights the crucial role that human forensic investigators play in mitigating these risks. Through adaptive decision-making, ethical reasoning, and contextual understanding, human investigators effectively identify subtle anomalies and threats that may evade automated detection systems. To reinforce our findings, we conducted comprehensive reliability testing of forensic techniques using multiple cyber threat scenarios. These tests confirmed that while AI agents significantly improve the efficiency of routine analyses, human oversight remains crucial in ensuring accuracy and comprehensiveness of the results.

In this study, we propose a homotopy-inspired prompt obfuscation framework to enhance understanding of security and safety vulnerabilities in Large Language Models (LLMs). By systematically applying carefully engineered prompts, we demonstrate how latent model behaviors can be influenced in unexpected ways. Our experiments encompassed 15,732 prompts, including 10,000 high-priority cases, across LLama, Deepseek, KIMI for code generation, and Claude to verify. The results reveal critical insights into current LLM safeguards, highlighting the need for more robust defense mechanisms, reliable detection strategies, and improved resilience. Importantly, this work provides a principled framework for analyzing and mitigating potential weaknesses, with the goal of advancing safe, responsible, and trustworthy AI technologies.

As malware continues to become increasingly sophisticated, threatening, and evasive, malware detection systems must keep pace and become equally intelligent, powerful, and transparent. In this paper, we propose Assembly Flow Graph (AFG) to comprehensively represent the assembly flow of a binary executable as graph data. Importantly, AFG can be used to extract granular explanations needed to increase transparency for malware detection using Graph Neural Networks (GNNs). However, since AFGs may be large in practice, we also propose a Meta-Coarsening approach to improve computational tractability via graph reduction. To evaluate our proposed approach we consider several novel and existing metrics to quantify the granularity and quality of explanations. Lastly, we also consider several hyperparameters in our proposed Meta-Coarsening approach that can be used to control the final explanation size. We evaluate our proposed approach using the CIC-DGG-2025 dataset. Our results indicate that our proposed AFG and Meta-Coarsening approach can provide both increased explainability and inference performance at certain coarsening levels. However, most importantly, to the best of our knowledge, we are the first to consider granular explainability in malware detection using GNNs.

In the network security domain, due to practical issues – including imbalanced data and heterogeneous legitimate network traffic – adversarial attacks in machine learning-based NIDSs have been viewed as attack packets misclassified as benign. Due to this prevailing belief, the possibility of (maliciously) perturbed benign packets being misclassified as attack has been largely ignored. In this paper, we demonstrate that this is not only theoretically possible, but also a particular threat to NIDS. In particular, we uncover a practical cyberattack, FPR manipulation attack (FPA), especially targeting industrial IoT networks, where domain-specific knowledge of the widely used MQTT protocol is exploited and a systematic simple packet-level perturbation is performed to alter the labels of benign traffic samples without employing traditional gradient-based or non-gradient-based methods. The experimental evaluations demonstrate that this novel attack results in a success rate of 80.19% to 100%. In addition, while estimating impacts in the Security Operations Center, we observe that even a small fraction of false positive alerts, irrespective of different budget constraints and alert traffic intensities, can increase the delay of genuine alerts investigations up to 2 hr in a single day under normal operating conditions. Furthermore, a series of relevant statistical and XAI analyses is conducted to understand the key factors behind this remarkable success. Finally, we explore the effectiveness of the FPA packets to enhance models' robustness through adversarial training and investigate the changes in decision boundaries accordingly.

Recent European efforts around digital identity – the EUDI regulation and its OpenID architecture – aim high to provide an EU-wide authentication framework. However, its current technical and legislative architecture are based on a limited conceptualization of identity. None of the legal and technical texts involved explicitly define this central term; and their implicit model of the concept does not go beyond a digitalization of identity cards and similar documents. Based on several other standards, we therefore propose a deeper, explicit definition. Grounded in this definition, we identify several issues in the design of OpenID4VCI and OpenID4VP, and show that neither the functional requirements nor the non-functional advantages claimed by OpenID's new trust model surpasses equivalent existing solutions. Also the EUDI legislation itself cannot accommodate its promise of self-sovereign identity. In particular, we criticize the introduction of institutionalized trusted lists, and discuss their economical and political risks. Their potential to decline into an exclusory, recentralized ecosystem endangers the vision of a user-oriented identity management in which individuals are in charge. In anticipation of revisions to the EUDI regulations, we suggest several technical alternatives for the OpenID architecture, as well as paths for future research, addressing a heterogeneity of attestations and providers.

GitHub Actions is a widely used platform to automate the build and deployment of software projects through configurable workflows. As the platform's popularity grows, it also becomes a target of choice for software supply chain attacks. These attacks exploit excessive permissions, ambiguous versions or the absence of artifact integrity checks to compromise the workflows. In response to these attacks, several security scanners have emerged to help developers harden their workflows. In this paper, we perform the first systematic comparison of 9 GitHub Actions Workflows security scanners. We compare them regarding scope (which security weaknesses they target), detection capabilities (how many weaknesses they detect), accuracy (how well their findings match manually confirmed weaknesses), and performance (how long they take to scan a workflow). In order to compare the scanners on a common ground, we first establish a classification of 10 common security weaknesses that can be found in GitHub Actions Workflows. Then, we run the scanners against a curated set of 2,722 workflows. Our study reveals that static scanners for GitHub Actions Workflows are very diverse, with both general purpose and focused scanners. More importantly, we provide evidence that these scanners implement fundamentally different analysis strategies, leading to major gaps regarding the nature and the number of reported security weaknesses. Based on this empirical evidence we make actionable recommendations for developers to harden their GitHub Actions Workflows.

Smart contracts are tools with self-execution capabilities that provide enhanced security compared to traditional contracts; however, their immutability makes post-deployment fault correction extremely complex, highlighting the need for a verification layer prior to this stage. Although formalisms such as Contract Language (CL) enable logical analyses, they prove limited in attributing responsibilities within complex multilateral scenarios. This work presents a proof of concept using the Relativized Contract Language (RCL) and the RECALL tool for the specification and verification of a purchase and sale contract involving multiple agents. The study demonstrates the tool's capability to detect normative conflicts during the modeling phase. After correcting logical inconsistencies, the contract was translated into Solidity and functionally validated within the Remix IDE environment, confirming that prior formal verification is fundamental to ensuring the reliability and security of the final code.

This survey reviews the existing and envisioned security vulnerabilities and defense mechanisms relevant to Advanced Air Mobility (AAM) systems, with a focus on electric vertical takeoff and landing (eVTOL) aircraft. Drawing from vulnerabilities in the avionics in commercial aviation and the automated unmanned aerial systems (UAS), the paper presents a taxonomy of attacks, analyzes mitigation strategies, and proposes a secure system architecture tailored to the future AAM ecosystem. The paper also highlights key threat vectors, including Global Positioning System (GPS) jamming/spoofing, ATC radio frequency misuse, attacks on TCAS and ADS-B, possible backdoor via Electronic Flight Bag (EFB), new vulnerabilities introduced by aircraft automation and connectivity, and risks from flight management system (FMS) software, database and cloud services. Finally, this paper describes emerging defense techniques against these attacks, and open technical problems to address toward better defense mechanisms.

The rapid expansion of IoT deployments has intensified cybersecurity threats, notably Distributed Denial of Service (DDoS) attacks, characterized by increasingly sophisticated patterns. Leveraging Generative AI through On-Device Large Language Models (ODLLMs) provides a viable solution for real-time threat detection at the network edge, though limited computational resources present challenges for smaller ODLLMs. This paper introduces a novel detection framework that integrates Chain-of-Thought (CoT) reasoning with Retrieval-Augmented Generation (RAG), tailored specifically for IoT edge environments. We systematically evaluate compact ODLLMs, including LLaMA 3.2 (1B, 3B) and Gemma 3 (1B, 4B), using structured prompting and exemplar-driven reasoning strategies. Experimental results demonstrate substantial performance improvements with few-shot prompting, achieving macro-average F1 scores as high as 0.85. Our findings highlight the significant advantages of incorporating exemplar-based reasoning, underscoring that CoT and RAG approaches markedly enhance small ODLLMs' capabilities in accurately classifying complex network attacks under stringent resource constraints.

Large Language Models (LLMs) are increasingly deployed as multi-turn assistants and customized through instruction tuning with project-specific training components. This practice creates a supply-chain risk when organizations reuse third-party fine-tuning frameworks, trainer extensions, or outsourced training code: an adversary who subtly compromises the loss-computation component can inject malicious supervision during fine-tuning while leaving the stored training corpus, model architecture, and deployment interface unchanged. Existing LLM backdoors and defenses are largely prompt-centric, relying on lexical, syntactic, or semantic patterns in user inputs while overlooking structural signals in multi-turn conversations. We propose Turn-based Structural Trigger (TST), a prompt-free backdoor that uses dialogue turn position as its activation condition. TST exploits structural cues implicitly encoded by chat templates and is implanted without modifying the stored dialogue corpus. Its trigger is automatically present once the conversation reaches the attacker-specified turn, making activation independent of downstream user inputs and resistant to prompt filtering, sanitization, and paraphrasing. In our primary setting, the model behaves normally during early interactions and activates the attacker-defined behavior only after the conversation reaches the designated structural condition. Across four open-source LLM families, TST achieves an average Attack Success Rate of 98.10% on target turns and a Clean Rate of 99.96% on non-target turns, while retaining 97.78% of clean-model utility. These results identify dialogue structure as an overlooked attack surface and motivate structure-aware auditing beyond prompt inspection.

Vision-Language-Action (VLA) models are increasingly deployed in safety-critical robotic applications, yet their security vulnerabilities remain underexplored. We identify a fundamental security flaw in modern VLA systems: the combination of action chunking and delta pose representations creates an intra-chunk visual open-loop. This mechanism forces the robot to execute K-step action sequences, allowing per-step perturbations to accumulate through integration. We propose SILENTDRIFT, a stealthy black-box backdoor attack exploiting this vulnerability. Our method employs the Smootherstep function to construct perturbations with guaranteed C2 continuity, ensuring zero velocity and acceleration at trajectory boundaries to satisfy strict kinematic consistency constraints. Furthermore, our keyframe attack strategy selectively poisons only the critical approach phase, maximizing impact while minimizing trigger exposure. The resulting poisoned trajectories are visually indistinguishable from successful demonstrations. Evaluated on the LIBERO, SILENTDRIFT achieves a 93.2% Attack Success Rate with a poisoning rate under 2%, while maintaining a 95.3% Clean Task Success Rate.

We consider the storage problem in an asymmetric $X$-secure private information retrieval (A-XPIR) setting. The A-XPIR setting considers the $X$-secure PIR problem (XPIR) when a given arbitrary set of servers is communicating. We focus on the trade-off region between the average storage at the servers and the average download cost. In the case of $N=4$ servers and two non-overlapping sets of communicating servers with $K=2$ messages, we characterize the achievable region and show that the three main inequalities compared to the no-security case collapse to two inequalities in the asymmetric security case. In the general case, we derive bounds that need to be satisfied for the general achievable region for an arbitrary number of servers and messages. In addition, we provide the storage and retrieval scheme for the case of $N=4$ servers with $K=2$ messages and two non-overlapping sets of communicating servers, such that the messages are not replicated (in the sense of a coded version of each symbol) and at the same time achieve the optimal achievable rate for the case of replication. Finally, we derive the exact capacity for the case of asymmetric security and asymmetric collusion for $N=4$ servers, with the communication links $\{1,2\}$ and $\{3,4\}$, which splits the servers into two groups, i.e., $g=2$, and with the collusion links $\{1,3\}$, $\{2,4\}$, as $C=\frac{1}{3}$. More generally, we derive a capacity result for a certain family of asymmetric collusion and asymmetric security cases.

Model-sharing platforms, such as Hugging Face, ModelScope, and OpenCSG, have become central to modern machine learning development, enabling developers to share, load, and fine-tune pre-trained models with minimal effort. However, the flexibility of these ecosystems introduces a critical security concern: the execution of untrusted code during model loading (i.e., via trust_remote_code or trust_repo). In this work, we conduct the first large-scale empirical study of custom model loading practices across five major model-sharing platforms to assess their prevalence, associated risks, and developer perceptions. We first quantify the frequency with which models require custom code to function and identify those that execute arbitrary Python files during loading. We then apply three complementary static analysis tools: Bandit, CodeQL, and Semgrep, to detect security smells and potential vulnerabilities, categorizing our findings by CWE identifiers to provide a standardized risk taxonomy. We also use YARA to identify malicious patterns and payload signatures. In parallel, we systematically analyze the documentation, API design, and safety mechanisms of each platform to understand their mitigation strategies and enforcement levels. Finally, we conduct a qualitative analysis of over 600 developer discussions from GitHub, Hugging Face, and PyTorch Hub forums, as well as Stack Overflow, to capture community concerns and misconceptions regarding security and usability. Our findings reveal widespread reliance on unsafe defaults, uneven security enforcement across platforms, and persistent confusion among developers about the implications of executing remote code. We conclude with actionable recommendations for designing safer model-sharing infrastructures and striking a balance between usability and security in future AI ecosystems.

Adversary emulation tools facilitate scripting and automated execution of cyber attack chains, thereby reducing costs and manual expert effort required for security testing, cyber exercises, and intrusion detection research. However, due to the fact that existing tools typically rely on agents installed on target systems, they leave suspicious traces that make it easy to distinguish their activities from those of real human attackers. Moreover, these tools often lack relevant capabilities, such as handling of interactive prompts, and are unsuitable for emulating specific stages of the kill chain, such as initial access. This paper thus introduces AttackMate, an open-source attack scripting language and execution engine designed to mimic behavior patterns of actual attackers. We validate the tool in a case study covering common attack steps including privilege escalation, information gathering, and lateral movement. Our results indicate that log artifacts resulting from AttackMate's activities resemble those produced by human attackers more closely than those generated by standard adversary emulation tools.

Split Learning (SL) offers a framework for collaborative model training that respects data privacy by allowing participants to share the same dataset while maintaining distinct feature sets. However, SL is susceptible to backdoor attacks, in which malicious clients subtly alter their embeddings to insert hidden triggers that compromise the final trained model. To address this vulnerability, we introduce SecureSplit, a defense mechanism tailored to SL. SecureSplit applies a dimensionality transformation strategy to accentuate subtle differences between benign and poisoned embeddings, facilitating their separation. With this enhanced distinction, we develop an adaptive filtering approach that uses a majority-based voting scheme to remove contaminated embeddings while preserving clean ones. Rigorous experiments across four datasets (CIFAR-10, MNIST, CINIC-10, and ImageNette), five backdoor attack scenarios, and seven alternative defenses confirm the effectiveness of SecureSplit under various challenging conditions.

Machine learning models are increasingly served behind APIs. This renders private prediction, i.e., privatizing a model's outputs rather than its parameters, a natural privacy target: model outputs are lower-dimensional and far more stable to training-data changes than weights. While differential privacy (DP) cannot effectively exploit this as it calibrates noise to worst-case sensitivity that is intractable to bound for non-convex models, we argue that PAC privacy is a natural fit for private prediction. It is instance-based, and calibrates noise to a black-box function's empirical stability to control mutual-information (MI) leakage. The missing ingredient is efficient, adaptive composition. Serving predictions means answering a long stream of adaptively chosen queries from untrusted users; existing composition either fails under adaptivity, grows quadratically, or reverts to input-independent, DP-like noise. We close this gap with a new adversarial composition result via adaptive noise calibration and prove that MI accumulates only linearly under adaptive and adversarial querying. Experiments across modalities show that prediction stability enables high utility even at a tiny per-query budget: on CIFAR-10, we achieve 87.79% accuracy with a per-query MI budget of $2^{-32}$. This enables serving one million queries while provably bounding membership-inference success to 51.08% – the same guarantee as $(0.04, 10^{-5})$-DP. Further, in the presence of auxiliary public data, the large volume of PAC-private predictions enables us to distill a publishable model that can be queried without limit. Concretely, 210,000 private labels on an ImageNet subset distill into a student reaching 91.86% accuracy on CIFAR-10 with membership inference success bounded by 50.49%, comparable to $(0.02, 10^{-5})$-DP.

Modern operating systems provide powerful mandatory access control mechanisms, yet they largely reason about who executes code rather than how execution originates. As a result, processes launched remotely, locally, or by background services are often treated equivalently once privileges are obtained, complicating security reasoning and enabling post-compromise abuse of sensitive system interfaces. We introduce origin-aware mandatory access control (OAMAC), a kernel-level enforcement model that treats execution origin – such as physical user presence, remote access, or service execution – as a first-class security attribute. OAMAC mediates access to security-critical subsystems based on execution provenance rather than identity alone, enabling centralized governance over multiple attack surfaces while significantly reducing policy complexity. We present a deployable prototype implemented entirely using the Linux eBPF LSM framework, requiring no kernel modifications. OAMAC classifies execution origin using kernel-visible metadata, propagates origin across process creation, and enforces origin-aware policies on both sensitive filesystem interfaces and the kernel BPF control plane. Policies are maintained in kernel-resident eBPF maps and can be reconfigured at runtime via a minimal userspace tool. Our evaluation demonstrates that OAMAC effectively restricts common post-compromise actions available to remote attackers while preserving normal local administration and system stability. We argue that execution origin represents a missing abstraction in contemporary operating system security models, and that elevating it to a first-class concept enables practical attack surface reduction without requiring subsystem-specific expertise or heavyweight security frameworks.

In this paper, we provide a unified security framework with the goal of being able to evaluate and compare different (known and new) approaches for authentication and encryption using, for example, DNA-based schemes. Therefore, we introduce and rigorously define chemical functions which model chemical systems as noisy challenge–response primitives, and formalize the associated chemical function infrastructure. Building on the theory of physical functions, we rigorously define robustness, unclonability, and unpredictability for chemical functions in both finite and asymptotic regimes, and specify security games that capture the adversary's power and the security goals. We instantiate the framework with the data from two existing DNA-based constructions ("operable random DNA" and "Genomic Sequence Encryption") and derive quantitative bounds for robustness, unclonability, and unpredictability. These results place DNA-based chemical functions on a rigorous cryptographic footing, enabling principled design and comparison of chemically grounded authentication mechanisms. We demonstrate applications to in-product authentication and to shared key generation using standard extraction techniques, showing the significance of this work regarding real-world applications.

Large language models (LLMs) have shown strong capabilities in multi-step decision-making, and are increasingly integrated into real-world agentic applications. While existing safety evaluations focus on static interactions or refusal rates, it remains largely underexplored what level of threat these models pose once an adversary successfully assigns them a malicious role. To evaluate this agentic misuse risk, we propose VirtualCrime, a dynamic multi-agent sandbox framework that elicits and measures multi-step criminal behavior from LLMs. We design 40 diverse crime tasks within this framework, covering 11 realistic maps and 14 objectives such as theft, robbery, and kidnapping. We evaluate 8 strong LLMs and find that: (1) even with robust safety alignment, all evaluated models exhibit alarming agentic capabilities to generate and execute detailed multi-step criminal plans when placed in an adversarial red-teaming setting; (2) models frequently suffer from behavioral alignment failures, exhibiting a high propensity for over-harm by escalating to violent means far beyond what the task requires; (3) their most sophisticated criminal behaviors concentrate in social deception and team coordination rather than technical exploitation. Our work highlights a critical gap in current safety paradigms, emphasizing the urgent need for behavioral alignment in autonomous agent deployments.

Traditional paper-based document management has long posed challenges related to security, authenticity, and efficiency. Despite advances in digitalization, official documents remain vulnerable to forgery, loss, and unauthorized access. This thesis proposes a decentralized infrastructure for digital notarization, signing, and sharing of documents using blockchain technology. The research addresses key issues of transparency, immutability, and feasibility by defining system requirements, evaluating existing solutions, and proposing a novel architecture based on distributed systems. By combining cryptographic techniques with decentralized storage, this research contributes to the development of a more secure and efficient framework for managing official documents. The findings highlight the potential of blockchain-based digital notarization to streamline bureaucratic processes, mitigate security risks, and enhance user trust in digital document management.

Public blockchains provide no native mechanism to verify the legal identity behind a deployed smart contract, which blocks institutional adoption and compliance with EU regulations such as MiCA and AMLR. We present KYC Seal, the first protocol that extends the EU eIDAS trust infrastructure to Ethereum smart contracts by cryptographically binding them to Qualified Electronic Seals issued by Qualified Trust Service Providers (QTSPs). The protocol realizes the full eIDAS trust chain, from the European Commission's List of Trusted Lists through Member-State trusted lists and QTSP-signed X.509 certificates down to the individual smart contract, natively on-chain. An on-chain parser extracts identity fields directly from the QTSP-signed certificate bytes at registration. Both cryptographic verifications, the QTSP issuance signature and the certificate holder's seal signature, are performed once at registration and cached as on-chain state, reducing per-interaction seal verification to a pure state check. A new P-256 elliptic-curve precompile in Ethereum (deployed December 2025) makes these one-time cryptographic steps economical, enabling trustless on-chain verification of eIDAS identities without oracles or runtime intermediaries. A reference implementation, a formal security analysis, and a gas evaluation are the subject of forthcoming work.

In the current era of interconnected cyberspace, there is an adverse effect of ransomware on individuals, startups, and large companies. Cybercriminals hold digital assets till the demand for payment is made. The success of ransomware upsurged with the introduction of Ransomware as a Service(RaaS) franchise in the darknet market. Obfuscation and polymorphic nature of malware make them more difficult to identify by Antivirus system. Signature based intrusion detection is still on role suffering from the scarcity of RaaS packet signatures. We have analysed RaaS samples by network forensic approach to investigate on packet captures of benign and malicious network traffic. The behavior analysis of RaaS family Ransomwares, Ryuk and Gandcrab have been investigated to classify the packets as suspicious, malicious, and non-malicious which further aid in generating RaaS packet signatures for early detection and mitigation of ransomwares belonging to RaaS family. More than 40% of packets are found malicious in this experiment. The proposed method is also verified by Virus Total API Approach. Further, the proposed approach is recommended for integration into honeypots in the present scenario to combat with data scarcity concerned with malware samples(RaaS). This data will be helpful in developing AI-based threat intelligence mechanisms. In turn enhance detection, prevention of threats, incident response and risk assessment.

Large language models (LLMs) are increasingly used for hardware and firmware code generation, but existing studies primarily evaluate functional correctness while largely overlooking security. However, LLM-generated code that appears functionally sound may embed security flaws which could induce catastrophic damages after deployment. This critical research gap motivates us to design a benchmark for assessing security awareness under realistic specifications. In this work, we introduce HardSecBench, a benchmark with 924 tasks spanning Verilog Register Transfer Level (RTL) and firmware-level C, covering 76 hardware-relevant Common Weakness Enumeration (CWE) entries. Each task includes a structured specification, a secure reference implementation, and executable tests. To automate artifact synthesis, we propose a multi-agent pipeline that decouples synthesis from verification and grounds evaluation in execution evidence, enabling reliable evaluation. We evaluate diverse LLMs and find that they often satisfy functional requirements while leaving security risks. We also find that security results vary with prompting. These findings highlight pressing challenges and offer actionable insights for future advancements in LLM-assisted hardware design. Our data and code are available at https://github.com/chenqirui2002/HardSecBench.

Robust reversible watermarking in encrypted images (RRWEI) faces an inherent challenge in simultaneously achieving robustness, reversibility, and content privacy under severely constrained embedding capacity. Existing RRWEI schemes often exhibit limited robustness against noise, lossy compression, and cropping attacks due to insufficient redundancy in the encrypted domain. To address this challenge, this paper proposes a novel RRWEI framework that couples dual most significant bit-plane (dual-MSBs) embedding with spatial redundancy and error-correcting coding. By compressing prediction-error bit-planes, sufficient embedding space and auxiliary information for lossless reconstruction are reserved. The dual-MSBs are further reorganized using a spiral embedding strategy to distribute multiple redundant watermark copies across spatially dispersed regions, enhancing robustness against both noise and spatial loss.Experimental results on standard test images demonstrate that the proposed method consistently outperforms under evaluated settings robustness against Gaussian noise, JPEG compression, and diverse cropping attacks, while maintaining perfect reversibility and high embedding capacity. Compared with state-of-the-art RRWEI schemes, the proposed framework achieves substantially lower bit-error rates and more stable performance under a wide range of attack scenarios.