Papers

Despite rapid progress in deep learning-based image watermarking, the capacity of current robust methods remains limited to the scale of only a few hundred bits. Such plateauing progress raises the question: How far are we from the fundamental limits of image watermarking? To this end, we present an analysis that establishes upper bounds on the message-carrying capacity of images under PSNR and linear robustness constraints. Our results indicate theoretical capacities are orders of magnitude larger than what current models achieve. Our experiments show this gap between theoretical and empirical performance persists, even in minimal, easily analysable setups. This suggests a fundamental problem. As proof that larger capacities are indeed possible, we train ChunkySeal, a scaled-up version of VideoSeal, which increases capacity 4 times to 1024 bits, all while preserving image quality and robustness. These findings demonstrate modern methods have not yet saturated watermarking capacity, and that significant opportunities for architectural innovation and training strategies remain.

Malware proliferation is increasing at a tremendous rate, with hundreds of thousands of new samples identified daily. Manual investigation of such a vast amount of malware is an unrealistic, time-consuming, and overwhelming task. To cope with this volume, there is a clear need to develop specialized techniques and efficient tools for preliminary filtering that can group malware based on semantic similarity. In this paper, we propose TrapNet, a novel, scalable, and unsupervised framework for malware fingerprinting and grouping. TrapNet employs graph community detection techniques for malware fingerprinting and family attribution based on static analysis, as follows: (1) TrapNet detects packed binaries and unpacks them using known generic packer tools. (2) From each malware sample, it generates a digest that captures the underlying semantics. Since the digest must be dense, efficient, and suitable for similarity checking, we designed FloatHash (FH), a novel numerical fuzzy hashing technique that produces a short real-valued vector summarizing the underlying assembly items and their order. FH is based on applying Principal Component Analysis (PCA) to ordered assembly items (e.g., opcodes, function calls) extracted from the malware's assembly code. (3) Representing malware with short numerical vectors enables high-performance, large-scale similarity computation, which allows TrapNet to build a malware similarity network. (4) Finally, TrapNet employs state-of-the-art community detection algorithms to identify dense communities, which represent groups of malware with similar semantics. Our extensive evaluation of TrapNet demonstrates its effectiveness in terms of the coverage and purity of the detected communities, while also highlighting its runtime efficiency, which outperforms other state-of-the-art solutions.

Radio Frequency Fingerprinting (RFF) has evolved as an effective solution for authenticating devices by leveraging the unique imperfections in hardware components involved in the signal generation process. In this work, we propose a Convolutional Neural Network (CNN) based framework for detecting rogue devices and identifying genuine ones using softmax probability thresholding. We emulate an attack scenario in which adversaries attempt to mimic the RF characteristics of genuine devices by training a Generative Adversarial Network (GAN) using In-phase and Quadrature (IQ) samples from genuine devices. The proposed approach is verified using IQ samples collected from ten different ADALM-PLUTO Software Defined Radios (SDRs), with seven devices considered genuine, two as rogue, and one used for validation to determine the threshold.

We present Core Mondrian, a scalable extension of the Original Mondrian partition-based anonymization algorithm. A modular strategy layer supports k-anonymity, allowing new privacy models to be added easily. A hybrid recursive/queue execution engine exploits multi-core parallelism while maintaining deterministic output. Utility-preserving enhancements include NaN-pattern pre-partitioning, metric-driven cut scoring, and dynamic suppression budget management. Experiments on the 48k-record UCI ADULT dataset and synthetically scaled versions up to 1M records achieve lower Discernibility Metric scores than Original Mondrian for numeric quasi-identifier sets while parallel processing delivers up to 4x speedup vs. sequential Core Mondrian. Core Mondrian enables privacy-compliant equity analytics at production scale.

The proliferation of high-fidelity synthetic media, coupled with exploitable hardware vulnerabilities in conventional imaging pipelines, has precipitated a crisis of trust in digital content. Existing countermeasures, from post-hoc classifiers to software-based signing, fail to address the fundamental challenge of establishing an unbreakable link to reality at the moment of capture. This whitepaper introduces Signing Right Away (SRA), a comprehensive security architecture that guarantees the provenance of digital media from "silicon to silicon to signed file." SRA leverages a four-pillar security model-Confidentiality, Integrity, Authentication, and Replay Protection, akin to the MIPI Camera Security Framework (CSF), but also extends its scope beyond the internal data bus to the creation of a cryptographically sealed, C2PA-compliant final asset. By securing the entire imaging pipeline within a Trusted Execution Environment (TEE), SRA ensures that every captured image and video carries an immutable, verifiable proof of origin. This provides a foundational solution for industries reliant on trustworthy visual information, including journalism, legal evidence, and insurance. We present the SRA architecture, a detailed implementation roadmap informed by empirical prototyping, and a comparative analysis that positions SRA as the essential "last mile" in the chain of content trust.

We propose a system for marking sensitive or copyrighted texts to detect their use in fine-tuning large language models under black-box access with statistical guarantees. Our method builds digital “marks” using invisible Unicode characters organized into (“cue”, “reply”) pairs. During an audit, prompts containing only “cue” fragments are issued to trigger regurgitation of the corresponding “reply”, indicating document usage. To control false positives, we compare against held-out counterfactual marks and apply a ranking test, yielding a verifiable bound on the false positive rate. Empirically, we obtain a true positive rate of 96.7% at 0% false positive rate and reply regurgitation rates exceeding 28% per document with only 40 (4%) watermarked documents. The approach is minimally invasive, scalable across many sources, robust to standard processing pipelines, and achieves high detection power even when marked data is a small fraction of the fine-tuning corpus.

Developers consistently use version constraints to specify acceptable versions of the dependencies for their project. Pinning dependencies can reduce the likelihood of breaking changes, but comes with a cost of manually managing the replacement of outdated and vulnerable dependencies. On the other hand, floating can be used to automatically get bug fixes and security fixes, but comes with the risk of breaking changes. Security practitioners advocate pinning dependencies to prevent against software supply chain attacks, e.g., malicious package updates. However, since pinning is the tightest version constraint, pinning is the most likely to result in outdated dependencies. Nevertheless, how the likelihood of becoming outdated or vulnerable dependencies changes across version constraint types is unknown. The goal of this study is to aid developers in making an informed dependency version constraint choice by empirically evaluating the likelihood of dependencies becoming outdated or vulnerable across version constraint types at scale. In this study, we first identify the trends in dependency version constraint usage and the patterns of version constraint type changes made by developers in the npm, PyPI, and Cargo ecosystems. We then modeled the dependency state transitions using survival analysis and estimated how the likelihood of becoming outdated or vulnerable changes when using pinning as opposed to the rest of the version constraint types. We observe that among outdated and vulnerable dependencies, the most commonly used version constraint type is floating-minor, with pinning being the next most common. We also find that floating-major is the least likely to result in outdated and floating-minor is the least likely to result in vulnerable dependencies.

The rapid spread of misinformation on digital platforms threatens public discourse, emotional stability, and decision-making. While prior work has explored various adversarial attacks in misinformation detection, the specific transformations examined in this paper have not been systematically studied. In particular, we investigate language-switching across English, French, Spanish, Arabic, Hindi, and Chinese, followed by translation. We also study query length inflation preceding summarization and structural reformatting into multiple-choice questions. In this paper, we present a multilingual, multi-agent large language model framework with retrieval-augmented generation that can be deployed as a web plugin into online platforms. Our work underscores the importance of AI-driven misinformation detection in safeguarding online factual integrity against diverse attacks, while showcasing the feasibility of plugin-based deployment for real-world web applications.

Quantum computing is rapidly emerging as one of the most transformative technologies of our time. With the potential to tackle problems that remain intractable for even the most powerful classical supercomputers, quantum hardware has advanced at an extraordinary pace. Today, major platforms such as IBM Quantum, Amazon Braket, and Microsoft Azure provide cloud-based access to quantum processors, making them more widely available than ever before. While a promising technology, quantum computing is not magically immune to security threats. Much research has been done on post-quantum cryptography, which addresses how to protect classical computers from attackers using quantum computers. This article meanwhile introduces the dual idea of quantum computer security: how to protect quantum computers from security attacks.

Generative AI leaderboards are central to evaluating model capabilities, but remain vulnerable to manipulation. Among key adversarial objectives is rank manipulation, where an attacker must first deanonymize the models behind displayed outputs – a threat previously demonstrated and explored for large language models (LLMs). We show that this problem can be even more severe for text-to-image leaderboards, where deanonymization is markedly easier. Using over 150,000 generated images from 280 prompts and 19 diverse models spanning multiple organizations, architectures, and sizes, we demonstrate that simple real-time classification in CLIP embedding space identifies the generating model with high accuracy, even without prompt control or historical data. We further introduce a prompt-level separability metric and identify prompts that enable near-perfect deanonymization. Our results indicate that rank manipulation in text-to-image leaderboards is easier than previously recognized, underscoring the need for stronger defenses.

We present BATTLE for Bitcoin, a DoS-resilient dispute layer that secures optimistic bridges between Bitcoin and rollups or sidechains. Our design adapts the BATTLE tournament protocol to Bitcoin's UTXO model using BitVM-style FLEX components and garbled circuits with on-demand L1 security bonds. Disputes are resolved in logarithmic rounds while recycling rewards, keeping the honest asserter's minimum initial capital constant even under many permissionless challengers. The construction is fully contestable (challengers can supply higher-work counter-proofs) and relies only on standard timelocks and pre-signed transaction DAGs, without new opcodes. For $N$ operators, the protocol requires $O(N^2)$ pre-signed transactions, signatures, and message exchanges, yet remains practical at $N\!\gtrsim\!10^3$, enabling high decentralization.

LLM-based agents are now used throughout cybersecurity. While these agents facilitate powerful and autonomous security applications, their autonomy opens up new attack surfaces, and the security community is actively building defenses to secure them. Yet the literature on this subject has grown quickly and unevenly. Existing surveys treat applications, threats, and defenses in isolation, leaving no unified account of how an agent's capabilities, vulnerabilities, and countermeasures interconnect. In this work we present the first holistic survey of the agentic security landscape, structuring the field around the fundamental pillars of Applications, Threats and Defenses. We provide a comprehensive taxonomy of over 260 papers, explaining how agents are used in downstream cybersecurity applications, inherent threats to agentic systems, and countermeasures designed to protect them. In addition, we provide detailed pillar-specific and cross-cutting analyses that show the security-lifecycle coverage of agentic applications, comparison between red-teaming and blue-teaming agents, and the adversarial use of red-teaming applications. On the threat side, we analyze the entry points and agent-loop stages that attacks target, their specificity to the agentic setting, and the threat models they assume. On the defense side, we analyze the prevailing defense strategies, their cost and security trade-offs, and where in the agent lifecycle they are deployed. We further map which defenses cover which attack classes and chart trends in agent architecture, backbone model usage, data modality coverage, and the growth of attack and defense research over time. Taken together, these findings indicate that agentic systems are structurally fragile by default and that securing them will require defenses that span the full agent lifecycle rather than single-layer fixes.

A central challenge in data security is not just preventing theft, but detecting whether it has occurred. Classically, this is impossible because a perfect copy leaves no evidence. Quantum mechanics, on the other hand, forbids general duplication, opening up new possibilities. We introduce Proofs of No Intrusion, which enable a classical client to remotely test whether a quantum server has been hacked and the client's data stolen. Crucially, the test does not destroy the data being tested, avoiding the need to store a backup elsewhere. We define and construct proofs of no intrusion for ciphertexts assuming fully homomorphic encryption. Additionally, we show how to equip several constructions of unclonable primitives with proofs of non-intrusion, such as unclonable decryption keys and signature tokens. Conceptually, proofs of non-intrusion can be defined for essentially any unclonable primitive. At the heart of our techniques is a new method for non-destructively testing coset states with classical communication. It can be viewed as a non-destructive proof of knowledge of a measurement result of the coset state.

The Precision Time Protocol (PTP), standardized as IEEE 1588, provides sub-microsecond synchronization across distributed systems and underpins critical infrastructure in telecommunications, finance, power systems, and industrial automation. While prior work has extensively analyzed PTP's vulnerability to network-based attacks, prompting the development of cryptographic protections and anomaly detectors, these defenses presume an uncompromised host. In this paper, we identify and exploit a critical blind spot in current threat models: kernel-level adversaries operating from within the host running the PTP stack. We present the first systematic study of kernel-rooted attacks on PTP, demonstrating how privileged attackers can manipulate system time by corrupting key interfaces without altering PTP network traffic. We implement three attack primitives, constant offset, progressive skew, and random jitter, using in-kernel payloads, and evaluate their impact on the widely used ptp4l and phc2sys daemons. Our experiments reveal that these attacks can silently destabilize clock synchronization, bypassing existing PTP security extensions. These findings highlight the urgent need to reconsider host-level trust assumptions and integrate kernel integrity into the design of secure time synchronization systems.

The emulation of multi-step attacks attributed to advanced persistent threats is valuable for training defenders and evaluating defense tools. In this paper, we discuss the numerous challenges and desired attributes associated with such automation. Additionally, we introduce the use of Effects Language (EL), a visual programming language with graph-based operational semantics, as a solution to address many of these challenges and requirements. We formally define the execution semantics of EL, and prove important execution properties. Furthermore, we showcase the application of EL to codify attacks using an example from one of the publicly available attack scenarios. We also demonstrate how EL can be utilized to provide proof-of-attack of complex multi-step attacks. Our results highlight the improvements in time and resource efficiency achieved through the use of EL for repeatable automation.

In safety-critical software systems, cybersecurity activities become essential, with risk assessment being one of the most critical. In many software teams, cybersecurity experts are either entirely absent or represented by only a small number of specialists. As a result, the workload for these experts becomes high, and software engineers would need to conduct cybersecurity activities themselves. This creates a need for a tool to support cybersecurity experts and engineers in evaluating vulnerabilities and threats during the risk assessment process. This paper explores the potential of leveraging locally hosted large language models (LLMs) with retrieval-augmented generation to support cybersecurity risk assessment in the forestry domain while complying with data protection and privacy requirements that limit external data sharing. We performed a design science study involving 12 experts in interviews, interactive sessions, and a survey within a large-scale project. The results demonstrate that LLMs can assist cybersecurity experts by generating initial risk assessments, identifying threats, and providing redundancy checks. The results also highlight the necessity for human oversight to ensure accuracy and compliance. Despite trust concerns, experts were willing to utilize LLMs in specific evaluation and assistance roles, rather than solely relying on their generative capabilities. This study provides insights that encourage the use of LLM-based agents to support the risk assessment process of cyber-physical systems in safety-critical domains.

The no-cloning theorem can be used as a basis for quantum money constructions which guarantee unconditionally unforgeable currency. Existing schemes, however, either (i) require long-term quantum memory and quantum communication between the user and the bank in order to verify the validity of a bill or (ii) fail to protect user privacy due to the uniqueness of each bill issued by the bank, which can allow its usage to be tracked. We introduce a construction of single-use quantum money that gives users the ability to detect whether the issuing authority is tracking them, employing an auditing procedure for which we prove unconditional security. Bill validation is classical, and hence does not require long-term quantum memory or quantum communication, making the protocol relatively practical to deploy. We discuss potential applications beyond money, including anonymous one-time pads and voting.

Chen, Liu, and Zhandry [CLZ22] introduced the problems $S|LWE\rangle$ and $C|LWE\rangle$ as quantum analogues of the Learning with Errors problem, designed to construct quantum algorithms for the Inhomogeneous Short Integer Solution ($ISIS$) problem. Several later works have used this framework for constructing new quantum algorithms in specific cases. However, the general relation between all these problems is still unknown. In this paper, we investigate the equivalence between $S|LWE\rangle$ and $ISIS$. We present the first fully generic reduction from $ISIS$ to $S|LWE\rangle$, valid even in the presence of errors in the underlying algorithms. We then explore the reverse direction, introducing an inhomogeneous variant of $C|LWE\rangle$, denoted $IC|LWE\rangle$, and show that $IC|LWE\rangle$ reduces to $S|LWE\rangle$. Finally, we prove that, under certain recoverability conditions, an algorithm for $ISIS$ can be transformed into one for $S|LWE\rangle$. We instantiate this reverse reduction by tweaking a known algorithm for $(I)SIS_\infty$ in order to construct quantum algorithm for $S|LWE\rangle$ when the alphabet size q is a small power of 2, recovering some results of Bai et al. [BJK+ 25]. Our results thus clarify the landscape of reductions between $S|LWE\rangle$ and $ISIS$, and we show both their strong connection as well as the remaining barriers for showing full equivalence.

Large reasoning models (LRMs) with multi-step reasoning capabilities have shown remarkable problem-solving abilities, yet they exhibit concerning safety vulnerabilities that remain poorly understood. In this work, we investigate why safety alignment fails in reasoning models through a mechanistic interpretability lens. Using a linear probing approach to trace refusal intentions across token positions, we discover a striking phenomenon termed as refusal cliff: many poorly-aligned reasoning models correctly identify harmful prompts and maintain strong refusal intentions during their thinking process, but experience a sharp drop in refusal scores at the final tokens before output generation. This suggests that these models are not inherently unsafe; rather, their refusal intentions are systematically suppressed. Through causal intervention analysis, we identify a sparse set of attention heads that negatively contribute to refusal behavior. Ablating just 3% of these heads can reduce attack success rates below 10%. Building on these mechanistic insights, we propose Cliff-as-a-Judge, a novel data selection method that identifies training examples exhibiting the largest refusal cliff to efficiently repair reasoning models' safety alignment. This approach achieves comparable safety improvements using only 1.7% of the vanilla safety training data, demonstrating a less-is-more effect in safety alignment.

In the context of Byzantine consensus problems such as Byzantine broadcast (BB) and Byzantine agreement (BA), the good-case setting aims to study the minimal possible latency of a BB or BA protocol under certain favorable conditions, namely the designated leader being correct (for BB), or all parties having the same input value (for BA). We provide a full characterization of the feasibility and impossibility of good-case latency, for both BA and BB, in the synchronous sleepy model. Surprisingly to us, we find irrational resilience thresholds emerging: 2-round good-case BB is possible if and only if at all times, at least $\frac{1}{\varphi} \approx 0.618$ fraction of the active parties are correct, where $\varphi = \frac{1+\sqrt{5}}{2} \approx 1.618$ is the golden ratio; 1-round good-case BA is possible if and only if at least $\frac{1}{\sqrt{2}} \approx 0.707$ fraction of the active parties are correct.

Mobile healthcare (mHealth) applications promise convenient, continuous patient-provider interaction but also introduce severe and often underexamined security and privacy risks. We present an end-to-end audit of 272 Android mHealth apps from Google Play, combining permission forensics, static vulnerability analysis, and user review mining. Our multi-tool assessment with MobSF, RiskInDroid, and OWASP Mobile Audit revealed systemic weaknesses: 26.1% request fine-grained location without disclosure, 18.3% initiate calls silently, and 73 send SMS without notice. Nearly half (49.3%) still use deprecated SHA-1 encryption, 42 transmit unencrypted data, and 6 remain vulnerable to StrandHogg 2.0. Analysis of 2.56 million user reviews found 28.5% negative or neutral sentiment, with over 553,000 explicitly citing privacy intrusions, data misuse, or operational instability. These findings demonstrate the urgent need for enforceable permission transparency, automated pre-market security vetting, and systematic adoption of secure-by-design practices to protect Protected Health Information (PHI).

A sum-product network (SPN) is a graphical model that allows several types of probabilistic inference to be performed efficiently. In this paper, we propose a privacy-preserving protocol which tackles structure generation and parameter learning of SPNs. Additionally, we provide a protocol for private inference on SPNs, subsequent to training. To preserve the privacy of the participants, we derive our protocol based on secret sharing, which guarantees privacy in the honest-but-curious setting even when at most half of the parties cooperate to disclose the data. The protocol makes use of a forest of randomly generated SPNs, which is trained and weighted privately and can then be used for private inference on data points. Our experiments indicate that preserving the privacy of all participants does not decrease log-likelihood performance on both homogeneously and heterogeneously partitioned data. We furthermore show that our protocol's performance is comparable to current state-of-the-art SPN learners in homogeneously partitioned data settings. In terms of runtime and memory usage, we demonstrate that our implementation scales well when increasing the number of parties, comparing favorably to protocols for neural networks, when they are trained to reproduce the input-output behavior of SPNs.

Provenance in scientific workflows is essential for understand- ing and reproducing processes, while in business processes, it can ensure compliance and correctness and facilitates process mining. However, the provenance of process adaptations, especially modifications during execu- tion, remains insufficiently addressed. A review of the literature reveals a lack of systematic approaches for capturing provenance information about adaptive workflows/processes. To fill this gap, we propose the AdProv method for collecting, storing, retrieving, and visualizing prove- nance of runtime workflow adaptations. In addition to the definition of the AdProv method in terms of steps and concepts like change events, we also present an architecture for a Provenance Holder service that is essential for implementing the method. To ensure semantic consistency and interoperability we define a mapping to the ontology PROV Ontol- ogy (PROV-O). Additionally, we extend the XES standard with elements for adaptation logging. Our main contributions are the AdProv method and a comprehensive framework and its tool support for managing adap- tive workflow provenance, facilitating advanced provenance tracking and analysis for different application domains.

Phishing websites remain a persistent cybersecurity threat by mimicking legitimate sites to steal sensitive user information. Existing machine learning-based detection methods often rely on supervised learning with labeled data, which not only incurs substantial annotation costs but also limits adaptability to novel attack patterns. To address these challenges, we propose PhishSSL, a self-supervised contrastive learning framework that eliminates the need for labeled phishing data during training. PhishSSL combines hybrid tabular augmentation with adaptive feature attention to produce semantically consistent views and emphasize discriminative attributes. We evaluate PhishSSL on three phishing datasets with distinct feature compositions. Across all datasets, PhishSSL consistently outperforms unsupervised and self-supervised baselines, while ablation studies confirm the contribution of each component. Moreover, PhishSSL maintains robust performance despite the diversity of feature sets, highlighting its strong generalization and transferability. These results demonstrate that PhishSSL offers a promising solution for phishing website detection, particularly effective against evolving threats in dynamic Web environments.

We classify small binary bibraces, using the correspondence with alternating algebras over the field F2, up to dimension eight, also determining their isomorphism classes. These finite-dimensional algebras, defined by an alternating bilinear multiplication and nilpotency of class two, can be represented by subspaces of skew-symmetric matrices, with classification corresponding to GL(m, F_2)-orbits under congruence. Our approach combines theoretical invariants, such as rank sequences and the identification of primitive algebras, with computational methods implemented in Magma. These results also count the number of possible alternative operations that can be used in differential cryptanalysis.

Decentralized Autonomous Organizations (DAOs) aim to enable participatory governance, but in practice face challenges of voter apathy, concentration of voting power, and misaligned delegation. Existing delegation mechanisms often reinforce visibility biases, where a small set of highly ranked delegates accumulate disproportionate influence regardless of their alignment with the broader community. In this paper, we conduct an empirical study of delegation in DAO governance off-chain discussions from 14 DAO forums. We develop a methodology to link forum participants to on-chain addresses, extract governance interests using large language models, and compare these interests against delegates' historical behavior. Our analysis reveals that delegations are frequently misaligned with token holders' expressed priorities and that current ranking-based interfaces exacerbate power concentration. We argue that incorporating interest alignment into delegation processes could mitigate these imbalances and improve the representativeness of DAO decision-making. To support future research, we will release our dataset and code in a public repository.

Security measures are essential in the automotive industry to detect intrusions in-vehicle networks. However, developing a one-size-fits-all Intrusion Detection System (IDS) is challenging because each vehicle has unique data profiles. This is due to the complex and dynamic nature of the data generated by vehicles regarding their model, driving style, test environment, and firmware update. To address this issue, a universal IDS has been developed that can be applied to all types of vehicles without the need for customization. Unlike conventional IDSs, the universal IDS can adapt to evolving data security issues resulting from firmware updates. In this study, a new hybrid approach has been developed, combining Pearson correlation with deep learning techniques. This approach has been tested using data obtained from four distinct mechanical and electronic vehicles, including Tesla, Sonata, and two Kia models. The data has been combined into two frequency datasets, and wavelet transformation has been employed to convert them into the frequency domain, enhancing generalizability. Additionally, a statistical method based on independent rule-based systems using Pearson correlation has been utilized to improve system performance. The system has been compared with eight different IDSs, three of which utilize the universal approach, while the remaining five are based on conventional techniques. The accuracy of each system has been evaluated through benchmarking, and the results demonstrate that the hybrid system effectively detects intrusions in various vehicle models.

Decentralized applications (dApps) in Decentralized Finance (DeFi) face a fundamental tension between regulatory compliance requirements like Know Your Customer (KYC) and maintaining decentralization and privacy. Existing permissioned DeFi solutions often fail to adequately protect private attributes of dApp users and introduce implicit trust assumptions, undermining the blockchain's decentralization. Addressing these limitations, this paper presents a novel synthesis of Self-Sovereign Identity (SSI), Zero-Knowledge Proofs (ZKPs), and Attribute-Based Access Control to enable privacy-preserving on-chain permissioning based on decentralized policy decisions. We provide a comprehensive framework for permissioned dApps that aligns decentralized trust, privacy, and transparency, harmonizing blockchain principles with regulatory compliance. Our framework supports multiple proof types (equality, range, membership, and time-dependent) with efficient proof generation through a commit-and-prove scheme that moves credential authenticity verification outside the ZKP circuit. Experimental evaluation of our KYC-compliant DeFi implementation shows considerable performance improvement for different proof types compared to baseline approaches. We advance the state-of-the-art through a holistic approach, flexible proof mechanisms addressing diverse real-world requirements, and optimized proof generation enabling practical deployment.

The Five Safes is a framework used by national statistical offices (NSO) for assessing and managing the disclosure risk of data sharing. It can be understood as a specialization of a broader concept–contextual integrity–to the situation of statistical dissemination by an NSO. We demonstrate this by mapping the five parameters of contextual integrity onto the five dimensions of the Five Safes. We also discuss how each of these two theories can address weaknesses in the other, thereby strengthening them both.

Supply chain security is extremely important for modern applications running at scale in the cloud. In fact, they involve a large number of heterogeneous microservices that also include third-party software. As a result, security vulnerabilities are hard to identify and mitigate before they start being actively exploited by attackers. For this reason, governments have recently introduced cybersecurity regulations that require vendors to share a software bill of material (SBOM) with end users or regulators. An SBOM can be employed to identify the security vulnerabilities of a software component even without access to its source code, as long as it is accurate and interoperable across different tools. This work evaluates this issue through a comprehensive study of tools for SBOM generation and vulnerability scanning, including both open-source software and cloud services from major providers. We specifically target software containers and focus on operating system packages in Linux distributions that are widely used as base images due to their far-reaching security impact. Our findings show that the considered tools are largely incompatible, leading to inaccurate reporting and a large amount of undetected vulnerabilities. We uncover the SBOM confusion vulnerability, a byproduct of such fragmented ecosystem, where inconsistent formats prevent reliable vulnerability detection across tools.