Papers

Agent skills package instructions, executable code, and task-specific resources into reusable artifacts that agents can improve using execution feedback. The same mechanism also enables attackers to evolve malicious skills, making them more effective and less detectable. However, a candidate skill may pass pre-execution scanning yet fail to realize its target under runtime defenses, while a revision that repairs execution may introduce new scanner findings. We introduce SkillDRE, a fully automated framework for evolving complete malicious skill packages through a dual-stage feedback loop. Given a benign task and its associated skills, SkillDRE autonomously constructs and validates a task-conditioned malicious objective and a verifiable judge rule. It then holds both fixed while evolving the skill implementation, with preservation of legitimate task capability. SkillDRE combines scanner-guided evolution with runtime-guided refinement informed by execution outcomes observed under runtime defense. Each runtime-guided revision returns to the pre-execution stage for rescanning and further optimization before re-execution, forming a cross-stage closed loop. Evaluated on SkillsBench across four victim models, SkillDRE achieves an average attack success rate of 45.28%, exceeding the strongest baseline by 40.3%, while its final submitted skills receive no SkillScan findings and largely preserve benign-task performance. These results show that two-stage defense feedback can serve as a useful learning signal for adaptive red teaming and that evaluating either defense stage in isolation can miss the resulting attack capability. Codes is available at https://github.com/whfeLingYu/SkillDRE

The July 2026 intrusion into Hugging Face production infrastructure showed how reward hacking can become an external cybersecurity incident when a capable agent encounters weak containment boundaries. This study develops a probabilistic risk model linking five stages: reward hacking, containment escape, usable access, persistence, and failure of detection. A Monte Carlo simulation evaluates 100,000 runs under each of four control configurations. Input distributions represent explicit uncertainty and are used for comparative analysis rather than real-world frequency prediction. Under the stated assumptions, layered controls reduce simulated external-incident probability substantially more than network isolation or monitoring used alone, an ordering that holds under independent plus/minus 25% perturbation of every coefficient in the model across 300 draws. Sensitivity analysis shows that agent capability and weaknesses in monitoring, authorization, and credential control exert the greatest influence on modeled risk. Human temporal discounting and metric gaming provide a behavioral analogy for short-horizon optimization, but the study does not infer that AI agents experience gratification or human motivation. The results support treating cyber-capable agent evaluations as hostile security zones in which indirect egress, shared infrastructure, credentials, and evaluation artifacts must remain outside the agent's effective authority.

Debate distillation adapts weaker verifiers using multi-agent debate transcripts to improve their judgement in subsequent debates, but gains on monitored tasks do not establish reliability on related unmonitored tasks. We study epistemic reliability degradation, in which adaptation preserves monitored performance while reducing support for correct responses on hidden tasks. We consider an adversarial debater that manipulates debate arguments while defending the correct monitored response, and ask whether the resulting degradation merely reflects catastrophic forgetting and whether standard evaluation can detect it. To address these questions, we propose ER-Audit, a two-stage black-box auditing framework that compares frozen verifier checkpoints before and after adaptation, and introduce two evaluation benchmarks pairing monitored and hidden task prompts grounded in shared contexts. ER-Audit searches for counterexamples to non-degradation by evaluating semantically valid paraphrases and, if none is found, uses independent paraphrases for sequential hypothesis testing. We derive anytime-valid lower confidence bounds on the non-degradation probability, allowing data-dependent stopping within a finite budget. We further establish a common lower bound across fixed paraphrase distributions and extend it to distributions within a bounded total variation distance of their mixtures. Our experiments show that higher hidden-task accuracy can coexist with more counterexamples to non-degradation and lower non-degradation bounds. This divergence challenges explanations based solely on broad catastrophic forgetting and shows that auditing can uncover selective hidden-task degradation concealed by aggregate performance gains. Our code and benchmarks are available at https://github.com/CSIRO-CQS-AI-alignment-Team/Epistemic-Reliability-Auditor.

Long-tailed encrypted traffic classification requires accurate recognition of infrequent classes under limited computational budgets. We propose Curv-Tail, a lightweight, end-to-end packet–byte framework trained without a separate pretraining stage. Mixed-resolution tokenization preserves exact packet-length identities within a bounded range and coarsens larger values to limit the vocabulary. An auxiliary objective predicts observed length tokens from contextual packet features before pooling, encouraging length-token retention beyond flow-level supervision. Compact temporal encoders process packet sequences and directional byte patches, and Lorentz prototypes with a shared learnable curvature magnitude classify their fused representation. On NUDT-Mobile and DataCon-Website under natural class frequencies, Curv-Tail achieves three-seed mean Tail-F1 scores of 85.70% and 46.30%, exceeding the strongest evaluated baselines by 2.91 and 1.89 percentage points, respectively. In 300-class profiling on an RTX 4090 with FP32 and batch size 256, Curv-Tail uses 98.48% fewer parameters and achieves 10.2 times the batch inference throughput of MM4Flow.

Neural image watermarks can be forged by extracting watermark-bearing residuals from released images and transferring them to unrelated content. While prior work has demonstrated this vulnerability, what makes these residuals transferable remains poorly understood. We formalize this vulnerability with residual transferability (RT), a metric that quantifies how well watermark evidence remains decodable after transfer across unrelated images. Through comparative analyses and controlled interventions, we find that common training-side variations do not account for the large RT differences across watermarking systems; instead, architectural design plays a central role. By contrasting high- and low-RT systems and validating their architectural differences through controlled interventions, we identify two mechanisms that strengthen the dependence of watermark evidence on the cover image, thereby suppressing the residual transferability. These findings provide concrete design guidance for developing more forgery-resistant watermarking architectures. Complementarily, for existing watermarking systems where architectural redesign is impractical, we introduce CoverLock, a plug-and-play strategy for existing watermarking systems that strengthens such image dependence without architectural redesign. Across representative watermarking systems exhibiting high residual transferability, CoverLock achieves a more favorable security–robustness trade-off than both traditional handcrafted defenses and learned classifier-based defenses.

Secure Transformer inference protects sensitive inputs but incurs substantial cryptographic overhead, with nonlinear operations such as Softmax and GeLU becoming major bottlenecks. Existing compression methods reduce nonlinear complexity, sequence-dependent computation, or model structure through separately defined compression variables. Under aggressive compression, however, these independently optimized perturbations can accumulate: at a matched compression level, stacking representative approximation, token-pruning, and model-pruning methods reduces ViT-S accuracy from 80.20% to 76.41%. We introduce DegreeSpar, which formulates secure Transformer compression as structured sparsification over nonlinear polynomial degrees. Polynomial degree directly controls the cost of secure nonlinear evaluation, while computation-aligned zero-degree structures expose token-level and model-dimension computation as removable within the same optimization space. DegreeSpar further incorporates approximation-aware training for low-degree Softmax and GeLU, enabling aggressive degree reduction and creating the optimization headroom required for structured computation removal. Across vision and language Transformers, DegreeSpar consistently improves the accuracy-latency trade-off across model scales, tasks, and sequence lengths, achieving speedups from 2.29x to 6.63x over the corresponding baselines. Under the same network setting, DegreeSpar achieves 92.68% accuracy on BERT/SST-2 in 110.55 s, compared with 92.66% in 167.26 s for CipherPrune, the closest prior hybrid secure-inference approach. These results establish structured polynomial degree as an effective shared optimization space for secure Transformer compression.

Persistent memory enables LLM agents to reuse prior experience, but creates a new security boundary: what an agent may remember is not what it should act on. We expose an unauthorized control path where edited low-trust evidence is consolidated into persistent memory, retrieved on a clean task, and used to drive a protected action. Crucially, the adversary neither writes memory nor alters the task. We introduce Backchain Memory Attack (BMA), a grey-box, LLM-driven inverse-planning attack that reasons backward from the target action to the memory that would trigger it, then to the evidence edit that would form it. BMA has two phases: preparation uses resettable trials to localize the first failed link and build experience; execution uses the frozen experience to rank and commit candidate edits without feedback. We introduce the Pathway-Certified Attack Success Rate (Path-CASR) to separate memory-mediated from coincidental hits: registered memory must form, be retrieved, drive the target behavior, and pass matched-intervention checks. Across four substrates and three decision backbones, BMA achieves 18.8% Macro Path-CASR, compared with 13.4% for the strongest access-matched baseline. Of BMA's behavioral hits, 60.3% pass all registered pathway and intervention checks versus 36.7% for the baseline. Frozen BMA edits retain 78.0% of their certified effect on average across four held-out consolidation policies. Representative memory-side controls leave 11.0% Path-CASR, whereas provenance-bound authorization reduces it to 2.0% while preserving 92.1% legitimate-action success.

When a language-model audit finds no leak, what is needed to certify non-leakage? We study guarantees over a declared prompt domain under an executable leakage criterion and decoding rule. For general bounded polynomial-time evaluators, a supplied leaking execution is polynomial-time checkable, while leak existence is \NP-complete and deterministic certification is \coNP-complete. Exact stochastic certification is $\coNP^{\PP}$-complete at every fixed rational cutoff in $(0,1)$. Restricting the computation can change these bounds. For example, certification is in \coNP\ when all randomness is a terminal draw from an efficiently computed finite probability table. Attention models admit polynomial-time certification when local dependency windows of logarithmic length precede one global head, given deterministic decoding, fixed vocabulary, exact rational weighted means, a direct binary affine readout and finite-automaton prompt domains. A construction with two global layers instead makes certification \coNP-complete over template domains, with one head per layer, polynomial width, logarithmic precision and an inverse-polynomial logit margin. Planted-secret experiments measure what finite audits miss relative to complete references. Among 30 secret–model-state pairs that leak under greedy single-prompt execution on their secret's 4,096-prompt domain, uniformly selecting 256 recorded evaluations per pair misses every leak for an expected $41.06\%$ of these pairs. Batched and single-prompt checks disagree on one complete-domain decision among all 48 fine-tuned pairs, while a same-order repeat reproduces every single-prompt output. These results distinguish computational conditions for certification from the coverage and execution conditions needed to interpret a negative audit.

Identity assurance is the cost a digital system pays for dishonesty and uncertainty: it exists to make acts attributable when not everyone can be trusted at their word. A common way to pay that cost is flat maximum verification, asking each participant to meet a single high level of identification at entry, before any capability is exercised. Paid on everyone, it over-collects, excludes participants who cannot meet a bar they never needed to clear, taxes every interaction with the cost of the rarest high-risk case, and binds the strength of identification to the activity it unlocks. Existing frameworks compound this by fixing a small number of per-credential levels inside a single legal space and binding each verification to the institution that performed it, leaving cross-border reuse and the tension between data erasure and evidentiary retention unaddressed. This paper develops, as a design-science proposal, a tiered and reusable model of identity assurance for natural, juridical, and machine entities across jurisdictions. Reading the problem through systems theory, where a system changes only when an entity acts, the model holds the assurance state apart from the capability gate that consumes it, so that identity demand follows the act and the weight of its consequences rather than mere presence: a participant may take part with minimal disclosure and supply more only as an act requires. It comprises a typed entity taxonomy, a two-axis coordinate of disclosed assertion scope and source of information, jurisdiction as a time-indexed attribute of the entity, and reliance recorded as bitemporal, liability-allocated, point-in-time snapshots. Requirements are derived from anti-money-laundering, electronic-identity, and data-protection law, and the proposal is evaluated against flat maximum verification, per-credential level-of-assurance designs, and institutional reusable-KYC reliance.

Runtime assurance pairs a verified fallback with an untrusted controller and a switching monitor, and is the leading route to admitting learned policies into safety-relevant network control. Its guarantee rests on a condition it assumes rather than requires: that the monitor's estimation error is zero, or at worst stochastic with a characterisable rate. In mobile networks many measurements originate at untrusted endpoints, where the error is neither. We state this assurance precondition and prove it: at zero measurement noise the guarantee survives an adversary controlling $q$ channels if and only if the plant is $2q$-sparse observable with respect to the safety-relevant output. This is functional observability over attack supports, strictly weaker than full-state observability: on our topology a per-cell outage condition doubles the budget and one stated over total offered load triples it. The converse is a constructive attack making a safe and an unsafe trajectory observationally identical; it defeats every monitor on those channels, not one detector, at any noise level. As the trust split is static and known before deployment, confining the adversary to one family yields a confinement threshold, above which trusted channels alone resolve the state and all untrusted channels may be corrupt at once. Three uninfluenceable counters cross it here, lifting the budget from two to six and making placement the dominant lever. Below the precondition the monitor faces a frontier, not a dichotomy; a set-valued monitor is optimal and carries sufficiency into non-zero noise at the residual test's false-rejection rate. An adversary reaching the trigger radius owns the switch regardless. The same argument bounds any monitor reading agent messages rather than measurements. A rank test decides the budget and identification does not preserve it, so a published budget must name its margin floor.

Privacy protection now spans legal, technical, and managerial duties, and demand for privacy professionals is growing across sectors. However, little is known about how employers define these roles. We analyze 1,143 U.S. privacy job postings from LinkedIn and Indeed. We examine job titles, salaries, competencies, certifications, education, experience, regulatory references, and AI-related language by using rule-based text mining. We also apply Topic Modeling (BERTopic) to the same postings and identify 18 latent themes which we grouped them into four categories. Our findings show that privacy roles are hybrid and they combine legal knowledge, technical skills, and interpersonal competence. Artificial intelligence appears in more than half of postings, with AI language spread across compliance, legal, governance and security themes. Our research indicates that AI governance responsibilities are often embedded within existing privacy roles, contributing to the rise of hybrid positions alongside dedicated AI governance roles.

Open-weight tool-calling agents are adopted on evidence of merit, usually benchmark scores and a record of reliable use. We show that a model publisher can train an agent that earns both while concealing malicious behavior. Fine-tuned on a mixture of clean and poisoned conversations, our agents answer ordinary requests correctly; once the system date reaches a chosen year, they emit the correct tool call and, alongside it, one that exfiltrates the user's credentials. The exfiltration runs while the user-facing response mentions only the legitimate work. We call this attack SilentCall. Under the trigger, it fires on at least $99.6$% of requests, and no response ever mentions it. The attack is detectable by three distinct methods, which differ mainly in what a defender needs to run them. A runtime monitor that inspects each tool call before it executes requires no access to the model and catches every instance of the payload we tested at a 1.73% false-positive rate. High-temperature probing requires only the published weights. The weight-distribution audit requires training a benign model with the suspect's recipe, placing it within reach of model hubs but not ordinary users. Alignment benchmarks, by contrast, do not separate poisoned from benign models. SilentCall leaves no trace on standard benchmarks. As tool-using agents spread through the open-weight supply chain, trust in them should not rest on what a model says about its own actions. It has to come from inspecting those actions at runtime, auditing models where they are distributed, and treating tool access as a security surface in its own right.

We propose Compact Shielded CSV, a private client-side validation blockchain for peer-to-peer payments designed for the postquantum era. Upgrading existing blockchains to quantum-resistant cryptography substantially increases on-chain overhead. By keeping all large cryptographic artifacts off-chain, Compact Shielded CSV keeps a minimal on-chain footprint independent of the size of the underlying cryptographic proofs and signatures. For a single input transaction, the onchain footprint is just 3 hashes (3x32 bytes): a nullifier, a degriefer, and a commitment to the transaction. We introduce the degriefer - a novel mechanism that enforces ownership and prevents double-spending using only hash commitments, eliminating the need for on-chain signatures entirely. These properties make Compact Shielded CSV a promising foundation for private, scalable, post-quantum digital payments.

Transaction fee mechanisms (TFMs) allocate scarce block space in blockchains. Prior work focuses on incentive compatibility (IC): honest behavior maximizes the payoff of each strategic player. We introduce rational-harm proofness (RHP), which rules out deviations that harm honest participants without reducing the deviator's utility relative to honest behavior. IC alone does not exclude such costless harm: in a second-price auction, a losing bidder can raise the winner's payment without changing its own payoff. For finite block capacity k, we characterize TFMs satisfying RHP alongside user incentive compatibility (UIC) and miner incentive compatibility (MIC). In the plain model, where a single miner implements the auction, we prove a tight tetrilemma: no TFM simultaneously achieves positive miner revenue, UIC, MIC, and RHP against miner-user coalitions, while any three are jointly achievable. In the MPC-assisted model, where a committee of miners implements the auction using multi-party computation, we construct a randomized TFM with positive miner revenue satisfying UIC, MIC, and RHP against users, miners, and miner-user coalitions containing at most k users. Randomness is necessary to confirm transactions under congestion: any deterministic TFM satisfying UIC and user RHP in this model confirms no transactions when the number of submitted bids exceeds k. Finally, IC and RHP are incomparable for every strategic role in both models: mechanisms can satisfy either property without satisfying the other. These results extend transaction fee mechanism design beyond incentives to protection against costless harm.

In the modern era, ransomware attacks on critical healthcare organizations—like hospitals and insurers—are frequent, with impacts ranging from leaked private health information all the way to serious disruptions of urgent, time-sensitive clinical operations. Unfortunately, legal and economic incentives make it uncommon for hospitals to share basic information about these attacks, their scope, and the downstream impacts to patient care. In this paper, we explore the use of public social media posts—authored both by hospitals and by individuals—to garner more detailed insights about these attacks and their effects. We collect 1,628 Facebook and Reddit posts from 2018–2024, design and evaluate techniques to match post data to 212 ground-truth ransomware attacks, and conduct quantitative and qualitative analyses that explore the impacts such attacks have on critical hospital infrastructure, patient care, and providers. We conclude by discussing the promise and limitations of leveraging social data to study the impact of ransomware attacks and highlight areas of future research.

As Internet of Things (IoT) networks increasingly depend on machine learning for anomaly, malware, intrusion detection, and network monitoring, such systems have become attractive targets for evasion attacks. Evasion attacks pose a major security risk because an adversary intentionally modifies input data to mislead a trained model into producing incorrect predictions while evading detection. This study evaluates the impact of black-box evasion attacks on a cost-utility-based adversarial training defense strategy in an Online AutoML context for IoT networks. Specifically, evasion attacks were applied to online learners, including Hoeffding Tree (HT), Leveraging Bagging (LB), Streaming Random Patches (SRP), Hoeffding Adaptive Tree (HAT), and Adaptive Random Forest (ARF). By developing naive and adversarially trained (AT) versions of these online learners, we generated clean and adversarial accuracies for each model. The results show that the AT versions of LB and SRP performed best, achieving the highest adversarial accuracy (0.985) and high clean accuracy (0.993) at the highest cost budget of 1.00, with a maximum accuracy reduction of only 0.8%. Finally, drift detection was conducted using the Early Drift Detection Method (EDDM).

When an oilfield digital twin disagrees with its instruments, the operator must decide whether the cause is hardware degradation, harsh-weather effects, or malicious data manipulation. Existing digital-twin work appears to treat these causes separately: sensor-validation architectures target faults, and attack-focused twins have been evaluated on water-sector testbeds. We study joint cause attribution on a simulated four-well wellpad with 15 coupled instruments, legitimate operating transients, and weather. A physics-informed twin, identified from normal data only, produces analytical-redundancy residuals; windowed features feed a gradient-boosted classifier whose posterior drives an alarm gate with a fixed false-alarm rate and a cost-based decision rule that may defer to an analyst. On three independently generated sites, the classifier reaches macro-F1 of 0.818 on windows where the injected deviation is observable (0.723 when latent post-onset windows are included at the primary site). The physics twin accounts for essentially all of this: removing the data-driven twin changes macro-F1 by less than 0.01, whereas removing all twins drops it to about 0.58. Attacks are detected quickly (median 1.7 h) but attributed correctly at alarm time only 42% of the time, rising to 73% six hours later. A cost-aware policy that defers ambiguous cases had the lowest expected cost among all policies in every one of 144 cost and prior settings tested, sometimes by a small margin; the result depends on illustrative costs and on analysts resolving deferred cases. A twin-aware attacker was detected in 45% of episodes yet almost never attributed to attack. These results are conditional on the simulator's generative assumptions and have not been validated on field data.

Knowledge distillation (KD) enables model transfer across heterogeneous platforms and deployment environments, yet it exposes proprietary models to distillation-based theft, where an adversary extracts intellectual property (IP) by training a student model on teacher outputs. Current defenses, such as software watermarking or hardware-based access control, either fail to survive the distillation process or lack the granularity to identify the specific device responsible for a leak. Aiming to provide post-theft accountability and traceability, we propose a novel fingerprinting framework that superimposes device-specific Physical Unclonable Function (PUF) signatures onto teacher logits during distillation. By utilizing signatures derived from Ring Oscillator (RO) PUFs measured on a Xilinx Zynq-7020 FPGA, we ensure that any student model trained via KD inherits a unique, hardware-linked identity. Our framework is architecture-agnostic, enabling reliable identity inheritance across heterogeneous structures, including Convolutional Neural Networks, Vision Transformers, and encoders. To ensure robust attribution, we implement a two-stage recovery pipeline consisting of a neural decoder and Hamming-distance refinement, maintaining high detection accuracy even under noisy conditions. Furthermore, we introduce a multi-level logit encoding scheme to support large-scale device deployment. Experimental results demonstrate that the embedded fingerprints are resilient against common post-distillation modifications. These results establish a practical system-level approach for enabling hardware-linked model traceability in distributed AI deployment environments.

Deep reinforcement learning has become a prominent approach to autonomous cyber defense. Existing methods are predominantly model-free and consequently require extensive environment interaction. World models provide an alternative by learning predictive dynamics and optimizing policies through imagined trajectories, yielding substantial gains in sample efficiency in robotics and embodied control. Extending this paradigm to cybersecurity raises a fundamental question: what should constitute the "world" in a cyber world model? We introduce CyberWorld, a Dreamer-style world modeling framework that learns latent cyber dynamics from vector, graph, textual, and multimodal representations of the defended network. Across all four scoreable CyberWheel attack strategies, the graph-based CyberWorld variant exceeds a strategy-agnostic control after 3.6k-15.8k environment steps, compared with millions of steps required by model-free PPO. Across representation choices, graph structure provides greater robustness under topology-dependent attacks, while simpler representations remain competitive in overall performance. Among successful runs, the number of episodes required to reach the control remains approximately constant as network size increases from 15 to 100 hosts. These results establish learned cyber dynamics as a sample-efficient and scalable basis for autonomous defense, and identify world representation as a central design axis for robustness and scalability.

LoRAs are widely studied for adapting base text-to-image diffusion models. However, a backdoored LoRA can hide a backdoor: it behaves normally in most cases, but produces attacker-specified content (the backdoor target) when a hidden backdoor trigger appears in the input prompt. Detecting such backdoors before using an untrusted LoRA is important for the safety of LoRA adaptation. We present TokenScanner, a model-level vocabulary scanner for backdoor detection within a LoRA fine-tuned text-to-image diffusion model, aiming to discover the malicious trigger for trustworthy LoRA adaptation. The key observation is that backdoor trigger tokens that appear in a larger proportion of training prompts tend to induce more prominent token-specific LoRA responses than those induced by unrelated tokens. TokenScanner therefore scans the tokenizer vocabulary and measures token-wise LoRA responses in the U-Net and the text encoder. It uses these responses to detect backdoored LoRAs and rank candidate trigger tokens for subsequent testing of backdoor activation. Experiments on seven backdoor settings, comprising 840 backdoored LoRAs and 840 real-world benign test LoRAs, show that TokenScanner achieves 95.96% AUC and 96.55% TPR at an FPR of 10.95%. It also achieves 89.40% Hit@1 and 99.40% Hit@5 for trigger discovery across all seven settings.

Cloud-based Large language model (LLM) services create a network-level traffic side channel that can expose model, prompt, and task behavior despite encryption. From packet sizes, directions, timing, and burst structure alone, a passive local observer can infer the serving model, the user's prompt category, and the task executed by a collaborative multi-agent system. Yet current evidence is fragmented across separate datasets and settings, limiting reproducibility and comparison. We present, to our knowledge, the first unified measurement study and public benchmark of encrypted LLM traffic across both user–LLM and multi-agent executions. The large-scale benchmark contains 60,000 user–LLM interactions across 10 models and 6 prompt categories, plus 2,838 multi-agent executions covering 10 task categories and two coordination topologies. Using only encrypted packet metadata, we assess the risk of traffic analysis attack by characterizing traffic signatures, identifying the features most associated with leakage, and testing robustness under prompt reformulation, decoding-temperature changes, larger candidate model sets, and partial traffic observation. Model fingerprinting achieves 97.7% balanced accuracy, prompt-category fingerprinting reaches 76.7% mean accuracy, and multi-agent task fingerprinting achieves up to 90.7% accuracy. Prompt reformulation weakens but does not remove model-specific leakage, and task fingerprints remain detectable even from a single agent's traffic.

The Manufacturer Usage Description (MUD) standard allows IoT manufacturers to define expected network behaviors in a MUD file. This file can be translated into enforceable access-control policies, restricting compromised devices to operate solely through manufacturer-defined communication patterns. However, practical adoption of MUD depends on profiles that are accurate, complete, and maintainable. Existing approaches use traffic-based automation but require device deployment and prolonged monitoring, capturing only behavior exercised during observation. Rare, failure-triggered, or configuration-dependent communications may remain absent, producing incomplete policies that disrupt legitimate operation and offer limited insight into the software components responsible for each rule. We present AutoMUD, a source-code-driven tool that generates traceable MUD profiles for IoT devices from their firmware and software source code. AutoMUD combines static and syntactic extraction, retrieval-grounded language-model reasoning, and deterministic validation and compilation to recover the communication behavior characterizing an IoT device and translate eligible endpoints into policy rules. By analyzing code-level evidence, AutoMUD exposes rarely exercised and conditional communication paths, links every generated rule to its source-level provenance, and preserves excluded findings with explicit reasons for review. Our evaluation on a Linux-based repository demonstrates that AutoMUD recovers complete communication behavior, consolidates validated behavior into semantic endpoint groups, and generates structurally valid MUD profiles. Through a controlled semantic fault-injection campaign, we demonstrate that AutoMUD enables analysts to detect, localize, explain, and correct propagated errors, recovering policies semantically identical to their clean counterparts.

Leaked system prompts are often treated as windows into the hidden values of commercial language models, yet their composition is rarely studied at scale. We analyze a merged corpus of 407 leaked, reconstructed, or officially published system prompts from 62 vendors across four community collections, identifying 29 near-duplicate clusters covering 66 files. Operational content rather than ethical statements dominates the corpus; a deliberately simple block-level classifier assigns roughly 58% of classified words to tool/protocol and roughly 5% to safety policy, while the strictest rule-lines guard tool use and file safety over harmful content by an 11:1 margin. Literal text transfer concentrates in a small set of cross-vendor pairs. Prompts also carry measurable maintenance debt, with version chains turning over thousands of words per release. The evidence supports treating leaked prompts as operational specifications, closer to configuration files than value statements, and treats reuse and prompt rot as engineering and supply-chain concerns. Because most documents are adversarial in origin and the detectors are deliberately simple, all magnitudes are directional; we audit the main classifier's error modes.

Contrastive Language–Image Pretraining (CLIP) has emerged as a dominant vision backbone due to its strong transferability and zero-shot capabilities. However, recent studies reveal a critical vulnerability: embedding-space backdoor attacks. By poisoning only a tiny fraction of image–text pairs, adversaries can implant stealthy triggers that induce targeted shifts in CLIP's joint embedding space. Unlike conventional backdoors that manipulate classifier logits, these attacks corrupt representations directly, making them highly effective under extremely low poisoning ratios and difficult to detect. Existing defenses require access to model parameters, gradients, logits, or clean validation data—assumptions that rarely hold in realistic black-box deployments. Moreover, current black-box methods struggle to accurately localize small or out-of-distribution triggers. We propose CLIPGuard, a lightweight and fully black-box defense specifically designed to mitigate embedding-space backdoors in CLIP encoders. CLIPGuard identifies malicious regions by measuring segment-wise embedding perturbations and selectively purifies only suspicious segments via semantic inpainting, preserving benign visual content and alignment quality. Extensive experiments on STL-10, ImageNet, and diverse trigger families—including BadCLIP, BadNets, blended, patch-based, and typographic attacks—demonstrate that CLIPGuard reduces attack success rates to as low as 1.05% while maintaining clean accuracy up to 86.34%, consistently outperforming existing black-box defenses, including CleanCLIP and CleanerCLIP. Our code is available https://github.com/wsu-cyber-security-lab-ai/CLIPGuard.git

As large language model (LLM) inference becomes increasingly expensive, resource-consumption attacks pose a growing threat to model providers. Existing attacks typically amplify cost by inducing abnormally long or repetitive outputs on attacker-controlled or triggered requests, making them easier to detect and limiting their deployment-wide impact when benign traffic dominates. In this work, we uncover a previously overlooked token-level attack surface arising from the many-to-one mapping from token sequences to decoded text. Although standard LLMs predominantly generate the canonical token sequences induced by their tokenizers, the same text can also be represented by substantially longer non-canonical sequences. This representational flexibility exposes a new avenue for resource-consumption attacks: an attacker can train the model to favor such sequences, systematically increasing the number of autoregressive decoding steps without a proportional increase in visible response length. However, we empirically find that directly maximizing token fragmentation substantially degrades model utility, producing conspicuous answer-quality failures that undermine attack stealthiness. To address this challenge, we propose FragToken, a training-time framework that combines source-model self-distillation, capacity-aware filtering and budgeting, and BPE-Aligned Merging to induce fragmented generation under ordinary prompts while largely preserving model utility. We evaluate FragToken on four LLMs across three benchmarks. Across the four models, FragToken achieves a three-benchmark average token inflation ratio (TIR) ranging from 1.99 to 2.46, while causing only minor degradation in model utility. Our work reveals a covert LLM supply-chain threat that increases inference cost without requiring large volumes of attack requests while largely preserving utility.

The Simultaneous Authentication of Equals (SAE) protocol, introduced in WPA3, provides robust protection against offline dictionary attacks against a network Pre-Shared Key (PSK) and also protection of session keys from other people knowing that PSK. However, its high computational cost for the Password Element (PE) derivation makes Access Points (APs) vulnerable to CPU exhaustion Denial-of-Service (DoS) attacks. This paper proposes a resilient architectural modification to SAE. First, we introduce an asymmetric resource cost model that offloads the iterative discovery of cryptographic elements to the client, allowing the AP to maintain a fixed computational load during the handshake. To further mitigate brute-force attempts, we implement a mechanism based on a slow-path key derivation (with variable cost key derivation functions, such as PBKDF2 or Argon2), incorporating a deliberate processing delay on the supplicant. Finally, we introduce a ticket-based mechanism to facilitate efficient re-authentication for known devices, bypassing expensive exchanges while preserving system availability under adversarial conditions. Experimental results demonstrate that this architecture significantly mitigates DoS risks without compromising legitimate network access.

Federated Learning (FL) allows devices with private data to collaborate in training a shared model. We present a next-generation FL system based on Trusted Execution Environments (TEEs) that addresses operational challenges associated with earlier systems and provides externally verifiable central Differential Privacy (DP) guarantees for the first time while offering a better privacy-utility tradeoff. In our system, devices upload data encrypted with keys managed by a TEE-hosted Key Management Service (KMS). The uploaded data is cryptographically tied to a policy limiting the set of Python programs that may later process the data in server-side TEEs. External parties may inspect public transparency logs to observe the set of workloads allowed by these policies. Our experimental results show that the new system improves device coverage and favorably shifts privacy-utility curves by enabling collected data to be integrated into the server-side workload at a schedule that optimizes DP guarantees and is unaffected by device availability. Our new system has been productionized, enabling models for the Android Keyboard (Gboard) to be trained faster and achieve better accuracy under smaller, now externally verifiable privacy budgets in comparison to models trained using the prior system.

As lotteries and other high-stakes decentralized applications increasingly depend on unpredictable randomness for their operations, the lack of a secure and transparent on-chain random number generator that is verifiable by all participants remains a critical open problem. Various approaches to blockchain-based random number generation have emerged over the years, each with their own strengths and limitations, and have consistently been superseded as blockchain technology evolved. This paper surveys existing approaches to on-chain randomness and proposes a new platform that builds upon the well-known commit-and-reveal scheme while directly addressing its principal vulnerability, the last revealer attack, in which the final participant can withhold their reveal in order to bias or abort the output upon seeing an unfavorable result. We further compare this solution with prior approaches and evaluate its entropy properties. The proposed architecture combines a web-based front-end with a Solidity smart contract deployed on the Polygon 2.0 blockchain. Implemented and tested on the Amoy testnet, the prototype is low-cost and simple to deploy, providing a practical, accessible proof-of-concept for verifiable on-chain randomness.

Third-party libraries (TPLs) are widely used in Android apps, but their reuse can introduce security risks and interfere with downstream program analyses. Existing Android TPL detection approaches face two key limitations: their hand-crafted features are fragile under aggressive code transformations, and their whole-library matching strategies are ineffective when apps retain only part of a TPL. In this paper, we propose LibFan, a learning-based Android TPL detection approach based on context-aware functional modeling. It realizes this modeling through two complementary components: context-aware contrastive learning at the method level and functional partitioning at the library level. At the method level, it learns semantic representations through contrastive training while incorporating outgoing call relationships and class-level context, improving robustness to obfuscation, shrinking, and optimization. At the library level, it partitions each TPL into functionally coherent units and determines library presence using the best-matching partition, thereby accommodating partial library reuse. To evaluate LibFan, we construct a new benchmark comprising 200 apps and 46 vulnerable TPLs, with each app compiled under four transformation configurations. Under the most challenging R8 full mode, LibFan achieves F1 scores of 81.3% at the library level and 47.6% at the version level, representing relative improvements of 64.9% and 35.6% over the state of the art, respectively.

When a card number is compromised, an attacker may search for active numbers sharing its prefix. An issuer might respond by reissuing cards from heavily populated prefixes into less populated ones. We show that this intuitive count control can backfire. For fixed search regions, exposure weights, and total activity, we derive exactly when reducing the maximum prefix count increases the chance that a budgeted search finds an active number. In matched synthetic simulations over 50,000 candidates, targeted replacement meets the count limit in every run but raises supplied-12-digit-prefix discovery relative to equal-volume random replacement in three of twelve settings. Thus a lower prefix count does not by itself certify lower enumeration risk.