Currently, existing group signature schemes with user-controlled and sequential linkability (GS-UCSL) suffer from critical limitations: they lack post-quantum security, cannot efficiently revoke malicious signers, grant excessive tracing power to group managers (GM), and rely on heavy revocation lists or secure channels that incur high computation or communication overhead. To address these issues, we propose an efficient lattice-based conditional privacy-preserving GS-UCSL (LCGS-UCSL). Our scheme is built on lattice cryptography to achieve post-quantum security, and unifies implicit linkability, explicit linkability, and sequential linkability under user autonomous control. We design a polynomial-based revocation mechanism that eliminates revocation list verification and secure-channel token updates, enabling lightweight and privacy-preserving member revocation. To curb GM’s overreach, we introduce key-oblivious encryption to split users into traceable and non-traceable types without their awareness, and use cuckoo hashing to realize O(1) registry operations. We further integrate signature aggregation and non-interactive zero-knowledge proofs of knowledge to optimize batch verification and reduce communication overhead. We formally prove the scheme’s anonymity, traceability, Existential Unforgeability under Chosen-Message Attack, and non-frameability in the random oracle model, and validate its practicality via performance analysis. The results show that LCGS-UCSL achieves comprehensive functionality with competitive efficiency, filling the gap toward post-quantum secure GS-UCSL with efficient revocation and balanced privacy.
Schwaemm is a member of the Sparkle-suite, a family of lightweight symmetric cryptographic algorithms that advanced to finalist status in the NIST lightweight cryptography standardization process. In 2023, using differential-linear cryptanalysis, Xiong and Liu presented a practical 4-round distinguisher and a theoretical 4.5-round key-recovery attack on a variant of Schwaemm128-128 that did not adhere to the original round constants. Recently, Niu et al. introduced a dedicated time-memory trade-off framework to achieve key-recovery attacks on 4.5-round Schwaemm using 2.5-round differential-linear distinguishers. This paper revisits the differential-linear cryptanalysis of round-reduced Schwaemm, focusing on the original, round-constant-respecting versions. We identify effective 4-round differential-linear trails and establish practical 4-round distinguishers with complexity below 2^13.1 . Experimental results demonstrate 100 2^9.58 and 2^15.73 ) and memory footprints (constrained between 2^27 and 2^55 ), significantly outperforming prior works. Furthermore, by extracting bit-level algebraic equations, we demonstrate that partial key information can be directly recovered with extremely low time complexities. For the respective versions, we establish a total of 36, 29, 29, and 9 equations, with time complexities of only 2^15.73 , 2^14.94 , 2^14.94 , and 2^9.58 . Among these, 9, 7, 7, and 3 are strictly linear equations, allowing for the direct recovery of an equal number of key bits, which drastically prunes the search space and reduces the time complexity of the full key-recovery process. Also, we emphasize that these attacks in this paper do not compromise the security of full Schwaemm.
Abstract Existing multi-turn jailbreak evaluation methods increasingly rely on large language models (LLMs) as automated judges to reduce the cost and scalability limitations of human assessment. However, recent studies show that LLM-based evaluators can diverge from human judgments under adversarial strategies involving subtle linguistic and semantic variations, raising reliability concerns in safety-critical domains such as cybersecurity. To address this challenge, we propose Refusal Manifold Guard (RefusalGuard-M), an open-source semantic evaluation framework that constructs a semantic refusal manifold from human-validated refusal responses for assessing LLM jailbreak interactions, including multi-turn scenarios. RefusalGuard-M uses embedding-based geometric representations to measure deviations from refusal behavior, providing a lightweight, interpretable, and reproducible alternative to LLM-based judging. We evaluate the framework across AdvBench, HarmBench, and CyMulTenSet, covering diverse jailbreak strategies, linguistic transformations, and multi-turn scenarios. Results show that RefusalGuard-M achieves strong agreement with human annotations and comparable recall performance to GPT-based evaluators while adopting a conservative evaluation strategy that prioritizes the detection of harmful outputs. On CyMulTenSet, which evaluates past-tense reformulated multi-turn jailbreaks, RefusalGuard-M achieves up to 0.87 recall, compared with 0.86 for GPT-5 and 0.81 for GPT-4, and reduces inference overhead by up to 3.7 $$\times$$ × relative to embedding-based baselines. These findings demonstrate that semantic refusal representations provide an efficient and scalable approach for jailbreak evaluation, particularly in cybersecurity settings where minimizing missed harmful outputs is critical.
Network traffic classification (NTC) increasingly needs to go beyond predefined categories and identify traffic from previously unseen classes, namely the open-set recognition problem in NTC. This problem remains challenging because encrypted traffic limits the practicality of payload-dependent features, while existing methods still lack representation learning tailored to open-set discrimination. This study aims to develop a payload-free open-set NTC framework that constructs an embedding space suitable for both known-class classification and unseen-class rejection. To this end, we propose SepSpace, which models network flows using meta-feature sequences. SepSpace introduces supervised contrastive learning to construct a discriminative embedding space with compact intra-class structures and separated inter-class boundaries. Combined with a prototype-radius rejection mechanism, SepSpace achieves effective unseen-class rejection while maintaining known-class classification performance. Meanwhile, to comprehensively evaluate open-set performance, we introduce a class-holdout evaluation protocol and conduct experiments on four benchmark datasets together with the conventional cross-dataset protocol. Experimental results demonstrate that, under the class-holdout protocol, SepSpace achieves the highest average area under the receiver operating characteristic curve (AUROC) of 0.9015 and open-set classification rate (OSCR) of 0.6351 among the evaluated methods; under the cross-dataset protocol, it obtains a higher average AUROC of 0.9396. The implementation and datasets of SepSpace are available at https://github.com/BITbla/SepSpace .
Existing hyperchaotic medical image encryption schemes derive all chaotic initial conditions solely from statistical features of deep-learning-segmented anatomical regions, creating a structural chosen-plaintext-attack (CPA) vulnerability: images with similar segmentation statistics yield cipher images from the same keystream, resulting in a Cipher Cross Correlation (CCC) close to unity (experimentally measured at 0.9141). We introduce a semantically conditioned key schedule that binds deep-learning semantic features with cryptographic hash functions to initialise chaotic dynamical systems. As a concrete instantiation, U-Net segmentation statistics and a SHA-256 digest of the whole plaintext image are fused via a nonlinear tanh coupling. The cryptographic guarantees structural CPA resistance, full-image plaintext sensitivity, and an effective key space of 2^256 derive from the keyed hash component under standard hash assumptions (complexity-theoretic IND-CPA is not claimed), while the U-Net component provides semantic key conditioning, enabling content-aware deployment scenarios such as audit traceability. The framework is drop-in compatible with existing permutation-diffusion pipelines and generalises beyond U-Net and SHA-256 to other semantic extractors and hash primitives. Experiments on six medical imaging datasets across five modalities confirm near-ideal entropy ( ≈ 7.997 bits), NPCR ( ≈ 99.61% ), and UACI ( ≈ 33.47% ), and a reduction of more than three orders of magnitude in adversarial CCC relative to the base architecture, at under 2 ms of additional computation per 256×256 image.
Abstract SCAN-C is a lightweight block cipher featuring a hybrid Feistel-SPN structure designed for Controller Area Network (CAN) security. This work presents the first full-round differential cryptanalysis of SCAN-C, demonstrating that its unique XNOR-based key mixing provides no additional security margin. We also prove that XNOR-based mixing is structurally equivalent to standard XOR mixing, as the corresponding difference distribution tables (DDTs) differ only by a permutation of indices. Utilizing an SMT-based automated search, we identify optimal differential clusters, including a 9-round distinguisher with a probability of $$2^{-51.30}$$ 2 - 51.30 . By extending these results, we show that the complete internal key state can be recovered through the recovery of all 12 round keys, requiring around $$2^{55}$$ 2 55 chosen plaintexts and $$2^{69}$$ 2 69 encryptions. Our results confirm that the current 12-round specification is insufficient for security and is the first to show that differential cryptanalysis is equivalent under both XOR and XNOR difference definitions.
Biometric cryptosystems designed with revocability and session-level renewability must preserve template privacy and maintain reliable authentication despite the inherent variability of biometric data. A key unresolved issue in deep biometric protection is how the embedding should be formed from a network’s final representation for secure cryptographic key binding. This study presents a privacy-preserving face biometric framework in which deep embeddings are transformed into binary revocable templates and bound on the fly to cryptographic keys using a fuzzy commitment scheme with error-correcting codes. This work focuses on revocability and session-level renewability; achieving full ISO/IEC 24745 cancelability (specifically unlinkability and irreversibility) requires a keyed transform and salted-hashed storage, which we identify as essential future work. Holding the trained network and the binding pipeline fixed, three feature-aggregation strategies applied to the final representation are systematically compared, namely Global Average Pooling, Smoothed Flattening, and a dedicated Dense Feature Layer. The resulting templates are combined with Reed–Solomon and Extended Hamming coding schemes to analyse the trade-offs among discriminability, error tolerance, key length, and computational cost. Experiments conducted on the FEI Face Database show that the Dense Feature Layer (a learned non-linear projection) provides a markedly more suitable representation for secure key binding than the non-learned spatial-aggregation alternatives, achieving the best authentication performance among the tested alternatives. At the selected operating point under controlled-acquisition, frontal-face conditions, the protected system achieves a 0
Recent studies have demonstrated that deep neural networks (DNNs) are highly vulnerable to bit-flip attacks (BFAs), where modifying only a few bits in memory can severely degrade model performance. However, existing BFA techniques typically rely on exhaustive searches over the entire parameter space and overemphasize single-bit flipping capability, leading to limited effectiveness in the presence of modern defenses. To address this challenge, we propose CritiFlip, an efficient BFA framework based on critical-weight identification. CritiFlip locates critical parameters through gradient computation and random perturbation, and performs targeted bit flips on them. By constraining the search to less than 0.1
Abstract The rapid growth of multimodal content has introduced new challenges in cybersecurity, particularly in scenarios such as misinformation detection, multimedia forensics, and open-source intelligence. In these settings, verifying the consistency between textual descriptions and video content is critical, yet remains challenging due to noisy, incomplete, or even misleading multimodal signals. Text-to-video retrieval provides a important capability for cross-modal alignment. However, existing approaches often overlook a key issue: the usefulness of different modalities varies over time and depends on the query. In particular, audio signals can be informative in certain segments (e.g., speech) but misleading in others (e.g., background noise), making uniform fusion unreliable in security-critical scenarios. In this work, we propose Temporal Uncertainty-aware Retrieval (TUR), a unified framework that models text-video alignment from two complementary perspectives: temporal modality reliability and alignment uncertainty. TUR dynamically estimates the contribution of multimodal signals over time and adapts text representations according to cross-modal agreement, enabling more stable retrieval. Extensive experiments on MSR-VTT, DiDeMo, VATEX, and LSMDC demonstrate that TUR consistently outperforms prior methods. Further analysis shows that TUR achieves improved temporal grounding, more stable similarity estimation, and enhanced interpretability, which are desirable properties for security-sensitive applications.
Transport Layer Cryptography Protocol (TLCP) is a secure communication protocol developed in China, featuring a dual-certificate architecture and incorporating ShangMi cryptographic algorithms. It has been widely deployed in security-critical domains such as finance, government, and energy. Despite its practical significance, TLCP did not undergo comprehensive formal analysis during its standardization process, leaving potential design-level vulnerabilities insufficiently explored. Moreover, the advent of quantum computing poses fundamental challenges to the classical cryptographic primitives employed by TLCP, motivating the need for both systematic security evaluation and post-quantum enhancements. To address these gaps, we first construct the comprehensive formal model of TLCP, covering certificate-based and identity-based cipher suites as well as its distinctive dual-certificate mechanism, under a realistic threat model and security assumptions that capture both classical and quantum adversaries. Based on this model, we conduct an automated security analysis using ProVerif, identifying nine potential attack vectors and deriving five concrete mitigation recommendations. Finally, motivated by the analysis results and the limitations of incremental fixes against quantum threats, we propose KEMTLCP, a post-quantum secure variant of TLCP that leverages key encapsulation mechanisms (KEMs) for both key exchange and authentication while preserving TLCP’s architectural principles through a novel explicit authentication mechanism. We further provide a security proof for the core authentication mechanism, show that KEMTLCP effectively mitigates the majority of identified vulnerabilities through formal analysis, and evaluate its practical performance.
Abstract Encrypted web traffic and evolving Internet technologies pose an increasing challenge to network traffic analysis. However, existing traffic classification methods, though effective, require large labeled datasets and complex training. This makes sustaining them prohibitively expensive and difficult in real-world scenarios. To narrow this gap, we propose a novel training-free few-shot network traffic classification framework based on large language models (LLMs). By integrating meta-learning with LLMs, it reduces reliance on labeled data, eliminates task-specific training, and improves performance. Specifically, we first apply an efficient feature extraction method to extract features from traffic flows. We then design meta-tasks that combine task descriptions with textualized features to produce natural language meta-task formulations. Building on these meta-tasks, the LLM performs reasoning to carry out traffic classification. Finally, to mitigate hallucination in the LLM outputs, we exploit the temporal characteristics of network traffic and aggregate predictions over samples within a defined time window. Extensive experiments on three widely-used encrypted traffic datasets demonstrate that our proposed framework outperforms the state-of-the-art methods, achieving an average absolute improvement in F1 score of 9.75, 9.82, and 12.06 percentage points on the three datasets, respectively.
Existing threshold Pointcheval–Sanders (PS) signature schemes largely assume equal-weight participants and therefore cannot directly express heterogeneous signing authority in semi-trusted multi-organization settings. This paper presents a dealer-assisted weighted threshold PS signing framework by integrating PS signatures with a Chinese Remainder Theorem (CRT)-based weighted ramp secret sharing (WRSS) mechanism. The main technical contribution is a CRT-to-PS integration method that converts weight-dependent modular shares into exponents while preserving the standard PS verification equation in prime-order bilinear groups. The WRSS base modulus is instantiated as the same prime order used by the bilinear-group scalar field, removing any modular mismatch between CRT reconstruction and PS exponentiation. As a result, the bit-length of each private-key share is proportional to the participant’s weight, and any signing subset whose total weight reaches the reconstruction threshold can jointly generate a valid PS signature. Meanwhile, any coalition whose total weight is at most the privacy threshold learns no information about the master secret beyond what is implied by the public key. We analyze the construction in the semi-honest model and in the WRSS-hybrid model, where the WRSS subroutines are idealized, and establish correctness, unforgeability, and randomization-based non-linkability. We further quantify the incremental computation and communication overhead introduced by the CRT-based weighting layer. The analysis separates CRT arithmetic from the bounded overflow-selection cost and includes heterogeneous-weight parameter tests. The proposed scheme is intended for dealer-assisted, semi-trusted, regulated multi-organization applications, such as joint authorization and credential-issuance workflows with policy-defined participant weights.
Abstract Large Language Models (LLMs) are increasingly deployed as interconnected agentic swarms that leverage the Model Context Protocol (MCP) to invoke shared external tools, APIs, and databases. In these settings, conventional security strategies based on agent-to-agent airgapping can be insufficient because agents that never directly communicate may still cross-infect one another through shared infrastructure tools, producing a “Confused Deputy” cascade. This paper develops a formal mathematical framework for modeling this cross-layer contagion in MCP-enabled multi-agent swarms. We construct a coupled multiplex Microscopic Markov Chain Approach (MMCA) that simultaneously tracks the node-level probability flow across two layers: an Agent cognitive layer (governed by Susceptible–Exposed–Infected–Quarantined (SEIQ) dynamics) and a Tool infrastructure layer (governed by Susceptible–Infected–Susceptible (SIS) dynamics). Our contributions are threefold: (i) we formulate a coupled multiplex MMCA with asymmetric SEIQ–SIS dynamics across agent and tool layers; (ii) we derive an analytical characterization of systemic risk, including an epidemic-threshold approximation via Next-Generation Matrix analysis and a closed-form budget-allocation rule under an exponential defense-efficiency model; and (iii) across eight experiment groups on synthetic and empirical agent-layer topologies, we show that shared-tool coupling consistently amplifies contagion and that tool-side controls can dominate agent-side hardening under the modeled regime. For cybersecurity practice, the framework identifies when shared tool infrastructure can transform localized prompt-injection events into system-level risk even under direct agent-to-agent isolation.
Smart home platforms predominantly adopt the broker-mediated publish/subscribe model (e.g., MQTT) for seamless device-app coordination. However, this architecture introduces a fundamental privacy-functionality conflict: the broker requires plaintext metadata (topic strings) for message routing, which inadvertently exposes fine-grained user behavioral patterns to semi-trusted service providers. Furthermore, existing systems lack rigorous cryptographic enforcement for app permissions, leaving the ecosystem vulnerable to over-privileged or malicious apps. While conventional Attribute-Based Encryption (ABE) provides fine-grained read-side access control, it cannot enforce writer-bound policies and remains computationally prohibitive for resource-constrained IoT nodes. In this paper, we propose SHERLOC, a practical and privacy-preserving framework that reconciles secure message routing with fine-grained privilege control. SHERLOC introduces two core primitives: (1) Secret Queue Telemetry Transport (SQTT), which leverages a novel trapdoor-based matching mechanism to support multi-level wildcard routing while ensuring topic indistinguishability and resistance against inside keyword-guessing attacks (IKGA); and (2) Outsourced Inner-Product Access Control Encryption (OS-IPACE), an attribute-hiding scheme that enforces dual no-read and no-write security for apps by offloading intensive pairing operations to a local hub without compromising data secrecy. We provide formal security proofs reducing SHERLOC’s privacy guarantees to the SXDH assumption. Experimental results from a full-scale prototype, comprising ESP32-based devices and Android apps, demonstrate that SHERLOC incurs millisecond-level latency and maintains compatibility with legacy MQTT brokers without altering protocol semantics, making it a robust and deployable solution for modern smart home environments.
Backdoor attacks pose a serious threat to Internet-of-Things (IoT) federated learning. In IoT deployments, pronounced non-independent and identically distributed (non-IID) data heterogeneity causes benign client updates to exhibit substantial variability across devices. Meanwhile, the physical exposure of IoT devices increases the risk of large-scale compromise and elevated malicious participation. Such variability allows poisoned updates to blend into natural fluctuations, rendering many robust aggregation and detection-based defenses unreliable. We propose FedDSG, a server-side defense that combines a semantic bias filter and a geometric direction constraint to counter backdoor manipulation. FedDSG first extracts a novel scale-invariant semantic cue from the last-layer bias of client updates to identify abnormal target-class reinforcement, staying effective even when benign bias patterns differ substantially across clients. The remaining updates are then constrained using a reference derived from a small trusted anchor set, limiting adversarial drift. This sequential design links semantic cues with geometric structure, where the former removes clearly suspicious updates and the latter stabilizes the residual ones, preventing misdetection-induced drift amplification while avoiding distortion of benign updates. The method does not alter client behavior or communication and adds minimal server-side overhead. Extensive experiments on MNIST, Fashion-MNIST, CIFAR-10, and SVHN under non-IID distributions with high malicious participation demonstrate the robustness of FedDSG. It reduces the attack success rate to 0.003, 0.006, 0.007, and 0.091, respectively, with only marginal accuracy loss and consistently achieves the highest Overall Performance Score (OPS), reflecting a superior trade-off between robustness and accuracy. Code and data availability information is provided in the Availability of data and materials section.
Combining traditional oil infrastructure with distributed renewable energy resources (DERs), including critical grid-edge assets like electric vehicle (EV) charging infrastructure that link transportation and power systems, has created hybrid energy systems that are more connected and digitally complex. These systems integrate legacy operational technologies (OT), such as SCADA, with decentralized, IoT-enabled assets that often rely on cloud-based analytics and remote connectivity. This conjunction introduces cybersecurity risks by linking historically isolated OT environments to modern, internet-exposed components, creating more entry points, inconsistent security baselines, and new attack surfaces. The resulting vulnerabilities extend beyond traditional IT-centric threat models. Conventional cybersecurity strategies, which prioritize technical controls in isolation, often fail to address systemic risks stemming from institutional fragmentation, regulatory gaps, and third-party dependencies. This study introduces a Socio-Technical Resilience Framework grounded in the Systems-Theoretic Accident Model and Processes (STAMP) and Socio-Technical Systems (STS) theory to access and mitigate cyber risks in hybrid grids. Through comparative case studies of the Colonial Pipeline ransomware event and cyber disruptions in European DER infrastructure, the paper finds that fragmented coordination, isolated threat intelligence, and weak human-system integration significantly amplify cyber impacts that undermine static, perimeter-based defense models. The study contributes a layered framework for cyber resilience that operates across three domains: (1) technical (via zero-trust architecture and distributed anomaly detection), (2) organizational (through shared situational awareness and cognitive decision-support tools), and (3) governance (by leveraging federated threat intelligence and regulatory harmonization). This multi-domain approach enhances both operational flexibility and long-term sustainability. Ultimately, the paper demonstrates that enduring cyber resilience in hybrid energy systems requires more than patching vulnerabilities. It demands a systemic rethinking of control, coordination, and design. The framework and findings offer a forward-looking roadmap for securing the energy sector against evolving threats in an era of distributed complexity.
Abstract Rust has garnered increasing attention as a programming language known for its high security and efficiency. However, new evidence suggests that attackers are now using Rust to develop novel ransomware variants, aiming to evade detection tools and maximize destructive capabilities. In this paper, we construct a real-world dataset consisting of 1823 Rust-based ransomware samples and conduct the empirical study that analyzes their behavioral patterns, revealing six attack features and two evasion mechanisms. Extensive observations show that Rust-based ransomware exhibits distinct behaviors compared to benign software, including semantically fragmented multi-process encryption and unique encryption patterns. Based on these insights, we design a lightweight and real-time detection method called RTCatch to counter evolving Rust-based ransomware attacks. The core of RTCatch lies in process correlation, which aggregates ransomware features across correlated processes, effectively addressing the challenge of feature dispersion caused by multi-process encryption. Experimental results demonstrate that RTCatch achieves a recall of 98.68% on the constructed dataset and performs rapid inference in 253 ms with a maximum response time of 2.97 s while incurring minimal computational and storage overhead. Furthermore, large-scale tests on 873 traditional ransomware samples show that RTCatch maintains consistent detection performance, indicating its robustness across different ransomware families.
