Papers

Conformance checking, one of the main process mining operations, aims to identify discrepancies between a process model and an event log. The model represents the expected behaviour, whereas the event log represents the actual process behaviour as captured in information systems' records. Traditionally, the process model and the event log are both accessible to the business analyst performing the conformance checking. However, in some contexts the log's owner may want to protect critical or sensitive information in the log and still check its conformance with respect to a model belonging to another party. In this paper, we propose a secure approach to conformance checking based on the well-known token-based replay algorithm and homomorphic encryption. An evaluation is performed using a synthetic log, showing the practicality of the proposed technique.

AI agents are autonomous entities that can be instantiated on demand, migrate across platforms, and interact with other agents or services without continuous human supervision. In such environments, identity is critical for establishing reliable interaction semantics among agents that may lack prior trust relationships. However, existing identity and access management mechanisms are designed for human users or static machines, assuming centralized enrollment, persistent identifiers, and stable execution contexts. These assumptions do not hold for AI agents, whose identities are self-managed, short-lived, and tightly coupled with their execution state and capabilities. We study the problem of identity authentication and state verification for AI agents and identify three challenges: (1) supporting self-managed identities for autonomously created agents, (2) enabling authentication under large-scale, concurrent interactions, and (3) verifying agents' dynamic execution state, such as whether their context and capabilities remain valid at interaction time. To address these challenges, we present AgentDID, a decentralized framework for identity authentication and state verification. AgentDID leverages decentralized identifiers (DIDs) and verifiable credentials (VCs), enabling agents to manage their own identities and authenticate across systems without centralized control. To address the limitations of static credential-based approaches, AgentDID introduces a challenge-response mechanism that allows verifiers to validate an agent's execution conditions at interaction time. We implement AgentDID in compliance with W3C standards and evaluate it through throughput experiments with multiple concurrent agents. Results show that the system achieves scalable identity authentication and state verification, demonstrating its potential to support large populations of AI agents.

We present MGTEVAL, an extensible platform for systematic evaluation of Machine-Generated Text (MGT) detectors. Despite rapid progress in MGT detection, existing evaluations are often fragmented across datasets, preprocessing, attacks, and metrics, making results hard to compare and reproduce. MGTEVAL organizes the workflow into four components: Dataset Building, Dataset Attack, Detector Training, and Performance Evaluation. It supports constructing custom benchmarks by generating MGT with configurable LLMs, applying 12 text attacks to test sets, training detectors via a unified interface, and reporting effectiveness, robustness, and efficiency. The platform provides both command-line and Web-based interfaces for user-friendly experimentation without code rewriting.

Agent Skills package SKILL.md files, scripts, reference documents, and repository context into reusable capability units, turning pre-load auditing from single-prompt filtering into cross-file security review. Existing guardrails often flag risk but recover malicious intent inconsistently under semantics-preserving rewrites. This paper formulates pre-load auditing for untrusted Agent Skills as a robust three-way classification task and introduces SkillGuard-Robust, which combines role-aware evidence extraction, selective semantic verification, and consistency-preserving adjudication. We evaluate SkillGuard-Robust on SkillGuardBench and two public-ecosystem extensions through five large evaluation views ranging from 254 to 404 packages. On the 404-package held-out aggregate, SkillGuard-Robust reaches 97.30% overall exact match, 98.33% malicious-risk recall, and 98.89% attack exact consistency. On the 254-package external-ecosystem view, it reaches 99.66%, 100.00%, and 100.00%, respectively. These results support a bounded conclusion: factorized package auditing materially improves frozen and public-ecosystem robustness, while harsher external-source transfer remains an open challenge.

Integrated sensing and communication (ISAC) combines sensing and communication within a shared system framework by using the same transmitted signal for both objectives. ISAC can improve the efficiency of spectrum and hardware use but also gives rise to new security challenges, as users associated with one function may need to be prevented from inferring information related to the other. This paper surveys information-theoretic approaches to secure ISAC with emphasis on formulations, performance metrics, and fundamental limits. We first review the information-theoretic ISAC models that underlie secure formulations. We then organize the secure ISAC literature according to the protected functionality and the adversary model, covering secure communication, sensing security, and active-adversary settings such as jamming. We also discuss formulations in which communication security and sensing security interact more directly, as well as their connections to privacy and covert communication. Throughout, we highlight the main modeling assumptions and the insights they provide on the tradeoffs among communication reliability, sensing performance, and security.

This paper presents C8s, a confidential computing architecture for Kubernetes that provides cryptographically rooted confidentiality, integrity, and verifiability guarantees for Kubernetes clusters from infrastructure operators. These guarantees are cryptographically provable to any independent third party verifier. The architecture is built on hardware Trusted Execution Environments (TEEs), specifically AMD SEV-SNP, Intel TDX, and NVIDIA Confidential Computing support, to establish an attestation-rooted trust boundary around confidential VMs. This design is compatible with managed Kubernetes services such as Amazon EKS, Google GKE, and Microsoft AKS, where the control plane cannot be attested. Under this boundary, three groups gain guarantees that are absent from conventional deployments. Data and artifact owners can deploy sensitive workloads and proprietary artifacts on third-party infrastructure without risking exfiltration. Compute providers can offer execution services without revealing workloads to cloud operators. End users can submit requests that remain opaque to all parties except the attested TEE processing them. Representative workloads include AI inference, securing AI model weights, and training or fine-tuning on sensitive data.

A decentralized autonomous organization (DAO) is a governing entity that empowers its stakeholders (i.e., users who hold one or more of its tokens) to manage blockchain-based protocols (i.e., smart contracts) collaboratively. The governance of a DAO is explicitly encoded in the DAO's governance contract, which defines how stakeholders participate in governance and how much influence (or voting power) they have in any decision. While decentralization and autonomy are the fundamental tenets of a DAO's design, empirical evidence suggests that in practice governance is often highly centralized. In this work, we study the designs and implementations of 48 public and actively used DAOs, with substantially large capital, deployed on Ethereum. We identify how three key governance mechanisms–token registration, staking, and delegation–originally introduced to improve security or participation, contribute to the concentration of voting power. Unlike prior work on centralization of voting power in specific DAOs, our findings reveal that these governance mechanisms of DAOs themselves systematically reinforce centralization. By elucidating the relationship between governance design and voting centralization, this work advances the understanding of DAO governance structures and highlights the inherent trade-offs between decentralization, security, and usability of DAOs.

The prevalence of biometric authentication has been on the rise due to its ease of use and elimination of weak passwords. To date, most biometric authentication systems have been designed for on-device authentication of the device owner (e.g., smartphones and laptops). Recently, biometric authentication systems have started to emerge that are designed to authenticate users against cloud databases storing representations of biometrics for large numbers of users (potentially millions), such as those facilitating biometric payments. However, the use of a large cloud database introduces a significant attack vector, as a breach of the database could lead to the compromise of all enrolled users' sensitive biometric data. Indeed, all such existing systems either do not adequately protect against such a breach, or are impractical to deploy and use due to their high computational overhead. In this work, we present a new biometric authentication system that provides provable security guarantees against data breaches, while remaining scalable and performant. To do so, we marry artificial intelligence with advanced cryptographic techniques in a novel fashion, providing several optimizations along the way. Our work is the first to show that real-world scalable privacy-preserving biometric authentication without auxiliary identifiers is feasible, and we believe that it will spur widespread industrial adoption and further research in this area.

Fully encrypted protocol-based tools (FEPs) are tools commonly used to circumvent censorship in restrictive regions, valued for their performance and security. However, in recent years, censors have been able to block them using an array of attacks based on passive traffic analysis and active probing. We propose Shaperd, an easily adoptable and real-time traffic shaper designed specifically to aid FEPs become more resilient to detection. Shaperd operates directly on packet contents in real time, using a novel constraint system to allow its users to generate traffic flows with any desired features. Our preliminary results reveal Shaperd introduces minimal overhead to the underlying system's throughput.

Collaborative blockchain ecosystems allow diverse groups to cooperate on tasks while providing properties such as decentralization and transaction security. We provide a model that uses a repository blockchain to manage hard forks within a collaborative system such that a single process (assuming that it has knowledge of the requirements of each fork) can access all of the blocks within the system. The repository blockchain replaces the need for Inter Blockchain Communication (IBC) within the ecosystem by navigating the networks. The resulting construction resembles a tree instead of a chain. A proof-of-concept implementation performs a depth-first search on the new structure.

Learned index structures achieve high performance by modeling the cumulative distribution function (CDF) of keys, but this reliance on data distributions introduces potential vulnerability to adversarial manipulation. Prior work has explored both static data poisoning and dynamic algorithmic complexity attacks (ACA), though evaluations are typically limited in scale or consider only one threat model. We present a systematic study of both attack paradigms on ALEX, a state-of-the-art dynamic learned index, under a unified and reproducible framework. Our evaluation scales to realistic workloads with up to 200K adversarial inserts and includes multiple SOSD datasets with diverse key distributions, as well as a real-key baseline to isolate adversarial effects. Our results show a clear separation between threat models. Static poisoning has minimal impact on lookup performance in ALEX under bulk-loaded settings, while dynamic ACA induces substantial degradation, with up to 2–2.8x slowdown in lookup throughput. However, attack effectiveness is highly dataset-dependent: dense key distributions limit adversarial leverage due to duplicate-heavy insertions and ALEX's localized structure. We highlight key evaluation considerations, including the need for control workloads and the mismatch between localized structural damage and global query metrics. These results show that robustness in learned indexes depends critically on the interaction between threat model, data distribution, and evaluation methodology.

The Controller Area Network (CAN) is a safety-critical in-vehicle communication protocol that lacks built-in security mechanisms, making intrusion detection essential. Existing approaches predominantly formulate CAN intrusion detection as a classification task, mapping complex traffic patterns to attack labels. However, this formulation abstracts away the temporal and relational structure of CAN traffic and misaligns with real-world forensic workflows, which require systematic reasoning about traffic behavior. To address this gap, we introduce CAN-QA, the first benchmark that reformulates CAN traffic analysis as a question-answering (QA) task. CAN-QA converts raw CAN logs into temporally segmented windows and applies deterministic rule-based templates to generate natural-language questions paired with automatically derived ground-truth answers. The resulting dataset comprises 33,128 QA pairs across 10 categories, each targeting distinct semantic and temporal properties of CAN traffic. Using CAN-QA, we evaluate large language models across both True/False and multiple-choice formats. Our results indicate that, although these models capture superficial statistical regularities, they struggle with temporal reasoning, multi-condition inference, and higher-level behavioral interpretation. Our code is available at https://github.com/Kriiiiss/CAN-QA.

Agentic systems now act with user secrets for APIs, messaging platforms, and cloud services. Today's agent runtimes typically implement authorization by exposure: enabling action means placing a reusable secret, or a reusable artifact derived from it, inside the runtime, so a transient prompt-injection or tool-side compromise becomes durable account compromise. Existing defenses cover adjacent pieces such as secret storage, scoped delegation, sender-constrained tokens, and runtime monitoring, but none states the combined requirement: an untrusted autonomous requester should be able to cause a user-authorized secret-backed operation without gaining reusable authority over it. We formalize this as the Agent Secret Use (ASU) problem and identify seven security properties any solution must satisfy, spanning authorization integrity and secret confidentiality. We propose the Secret-Use Delegation Protocol (SUDP), in which a requester proposes a canonical operation, the user authorizes it with a fresh authenticator-backed grant, and a custodian redeems the grant to perform the bounded use; reusable authority never crosses the requester boundary. We specialize SUDP for LLM-driven agents, where it applies whenever a tool call would exercise user-enrolled authority-bearing material. Under standard cryptographic assumptions, SUDP satisfies six of the seven properties; the seventh, runtime-memory confidentiality, is closed by composition with a hardware-rooted runtime. A reference implementation, available at https://github.com/xhyumiracle/sudp, adds sub-millisecond protocol overhead per mediated call.

The rapid rise of generative AI has made it easy to create convincing fake media at scale. In response, an industrial coalition has developed the Coalition for Content Provenance and Authenticity (C2PA), a system intended to provide verifiable provenance for digital content. Our research team conducted the first comprehensive, independent security analysis of C2PA. Our study includes the first formal-methods analysis of C2PA's core protocols. We find that the current C2PA specifications fail to achieve their claimed security goals. Furthermore, they also fail to achieve key additional goals, which all such provenance systems require for trustworthy deployment. As a result, C2PA may mislead users, platforms, and policymakers if relied upon prematurely. C2PA is a promising idea, but it should not yet be relied upon for high-stakes uses such as financial disclosures, journalism, or legal evidence.

Post-Quantum Cryptography (PQC) is a rapidly growing deployment challenge as cryptographically relevant quantum computers (CRQC) continue to advance, leaving traditional cryptographic algorithms used in X.509 vulnerable to attack. However, PQC introduces significant deployment challenges in real-world networks, with handshake sizes increasing from 5x to over 20x compared to classical algorithms. In this work, we evaluate the time to first byte (TTFB) under CDN-focused TLS conditions to characterize the latency cost of transitioning existing internet infrastructure to quantum-safe certificate schemes. We observe discrete increases in TTFB as certificate chain sizes exceed transport layer data flight limits. To isolate the impact of certificate chains, we evaluate both ECDSA and ML-DSA-based certificate schemes, generating similarly sized certificate chains through controlled addition of certificate extensions. We additionally examine how CDN properties such as session resumption, certificate size optimizations, and geographical distribution reduce latency penalties. We utilize Zeek-monitored TLS traffic through a High-Performance Computing System (NCSA) with terabyte network connectivity across the nation to quantify real-world session resumption rates. We compare CDN-driven size optimization with Merkle Tree Certificates (MTC) to examine how size reductions allow certificate chains to remain under the flight limit threshold. We find that MTC allows for 2x-3x increase in supportable certificate chain size, whereas CDN-based optimizations yield more limited reductions, supporting up to approximately 1.6x certificate chain size increase.

This report presents a comparative evaluation of DKnownAI Guard in AI agent security scenarios, benchmarked against three competing products: AWS Bedrock Guardrails, Azure Content Safety, and Lakera Guard. Using human annotation as the ground truth, we assess each guardrail's ability to detect two categories of risks: threats to the agent itself (e.g., instruction override, indirect injection, tool abuse) and requests intended to elicit harmful content (e.g., hate speech, pornography, violence). Evaluation results demonstrate that DKnownAI Guard achieves the highest recall rate at 96.5% and ranks first in true negative rate (TNR) at 90.4%, delivering the best overall performance among all evaluated guardrails.

Side Channel Analysis (SCA) relaxes the black-box assumption of conventional cryptanalysis by incorporating physical measurements acquired during cryptographic operations. Electro-magnetic (EM) emissions of a chip during computations often provide a very valuable source of side channel leakage. During the evaluation of a chip for electro-magnetic side channel emissions one needs to position an electro-magnetic probe in an advantageous position relative to the chip. Previous literature focused on hot-spot finding and to a lower extend repositioning. Trace augmentations have been considered to aid portability of profiling using one physical device and attacking another device. This paper focuses on training a single neural network using traces from multiple EM probe positions to detect leakage from a larger area over the attacked device. We provide dual evaluation of EM traces - from two completely independent labs - profiling on data from one lab and attacking traces from the other lab.

Barrett reduction is the nonlinear core of every practical NTT-based post-quantum cryptography implementation. Existing composition frameworks (ISW, t-SNI, PINI, DOM) address Boolean masking over GF(2); none provides a machine-checked characterization of Barrett's leakage under first-order arithmetic masking and the first-order probing model over prime fields. Building on our prior series, QANARY [15], partial-NTT-masking margins [14], algebraic foundations [16], and butterfly composition [18], we close this gap. We prove a trichotomy: for any $q > 0$ and shift $s$, the Barrett internal wire map $f_x(m) = ((x + 2^s - m) \bmod 2^s) \bmod q$ has preimage cardinality in $\{0, 1, 2\}$, never more. We call this the 1-Bit Barrier: max-multiplicity 2 implies at most 1 bit of min-entropy loss per internal wire, universal over all moduli. The count-zero cases, unreachable output values, reveal that actual leakage is often strictly less than 1 bit, making the bound conservative. We introduce PF-PINI (Prime-Field PINI): Barrett satisfies PF-PINI(2); the Cooley-Tukey butterfly satisfies PF-PINI(1). We observe (not yet proved) that with fresh inter-stage masking, the composed pipeline has max-multiplicity $\max(k_1, k_2)$, so the 1-Bit Barrier propagates. The trichotomy, the PF-PINI instantiations, and cardinality results are machine-checked in Lean 4 with Mathlib: 12 proved results, zero sorry, universal over all $q > 0$ (the min-entropy bound follows by standard definitions). Adams Bridge lacks fresh inter-stage masking, violating PF-PINI composition and explaining why Papers 1 [15] and 2 [14] found vulnerabilities. NIST IR 8547 recommends formal methods for PQC implementation validation. The 1-Bit Barrier provides the first universal machine-checked cardinality bound for masked Barrett reduction in ML-KEM (FIPS 203) and ML-DSA (FIPS 204), with a corresponding 1-bit leakage interpretation.

Autonomous AI agents extend large language models into full runtime systems that load skills, ingest external content, maintain memory, plan multi-step actions, and invoke privileged tools. In such systems, security failures rarely remain confined to a single interface; instead, they can propagate across initialization, input processing, memory, decision-making, and execution, often becoming apparent only when harmful effects materialize in the environment. This paper presents AgentWard, a lifecycle-oriented, defense-in-depth architecture that systematically organizes protection across these five stages. AgentWard integrates stage-specific, heterogeneous controls with cross-layer coordination, enabling threats to be intercepted along their propagation paths while safeguarding critical assets. We detail the design rationale and architecture of five coordinated protection layers, and implement a plugin-native prototype on OpenClaw to demonstrate practical feasibility. This perspective provides a concrete blueprint for structuring runtime security controls, managing trust propagation, and enforcing execution containment in autonomous AI agents. Our code is available at https://github.com/FIND-Lab/AgentWard .

Current cyber attribution approaches typically operate on a per-incident basis, leaving open whether aggregating evidence across campaigns improves adversary identification. We investigate whether cross-campaign attribution reduces ambiguity or whether structural limits persist under longitudinal data. We model adversary fingerprints as multi-dimensional feature vectors encoding behavioral, infrastructural, and temporal characteristics derived from covert beacon interactions. We introduce ARCANE (Attacker Re-identification via Cross-campaign Attribution Network), a probabilistic framework that aggregates passive telemetry across campaigns and organizations to construct persistent adversary fingerprints. These fingerprints are updated using a Bayesian belief network that integrates new evidence over time. A time-decayed confidence metric captures accumulated similarity across campaigns. Evaluation on a synthetic dataset of multiple threat profiles shows that intra-actor similarity consistently exceeds inter-actor similarity. However, separation between distinct actors remains limited due to shared operational practices among sophisticated adversaries. Results indicate that cross-campaign aggregation alone does not resolve attribution ambiguity. Performance is constrained by a structural ceiling in feature space, where inter-actor similarity remains high even without evasion. Attribution accuracy remains stable under increasing evasion, suggesting the main limitation is feature indistinguishability rather than adversarial adaptation. These findings highlight the need for additional signal classes, such as targeting patterns, temporal coordination, and infrastructure relationships, to improve attribution reliability.

Object detection (OD) is critical to real-world vision systems, yet existing backdoor attacks on detection transformers (DETRs) for OD tasks rely on patch-wise triggers optimized at fixed locations with minimal perturbations. Such attacks overlook that backdoor triggers in the real world may appear at different sizes, fields of view (FoVs), and locations in images, while minimal perturbations are difficult for cameras to capture, limiting attack practicality. We first observe that a patch-wise trigger in DETR delivers high attack effectiveness when activating the backdoor across neighboring locations, a phenomenon we term the trigger radiating effect (TRE). Meanwhile, inserting patch-wise triggers across multiple locations synergistically enhances TRE, resulting in high attack effectiveness across images. We propose DETOUR, a practical backdoor attack by using semantic triggers that are effective in real-world object detection systems. To ensure attack practicality, we rescale trigger patterns to different sizes and insert them at various predefined locations during backdoor training, enabling the model to recognize the trigger regardless of its spatial configurations. To address FoV variations in physical deployments, we extract the trigger pattern from a real-world object (e.g., a mug) captured under multiple FoVs and inject the trigger accordingly, promoting viewpoint-invariant backdoor activation and enhancing TRE across the entire image. As a result, the backdoor can be reliably activated under diverse FoVs and spatial configurations.

Large language models deployed at runtime can misbehave in ways that clean-data validation cannot anticipate: training-time backdoors lie dormant until triggered, jailbreaks subvert safety alignment, and prompt injections override the deployer's instructions. Existing runtime defenses address these threats one at a time and often assume a clean reference model, trigger knowledge, or editable weights, assumptions that rarely hold for opaque third-party artifacts. We introduce Layerwise Convergence Fingerprinting (LCF), a tuning-free runtime monitor that treats the inter-layer hidden-state trajectory as a health signal: LCF computes a diagonal Mahalanobis distance on every inter-layer difference, aggregates via Ledoit-Wolf shrinkage, and thresholds via leave-one-out calibration on 200 clean examples, with no reference model, trigger knowledge, or retraining. Evaluated on four architectures (Llama-3-8B, Qwen2.5-7B, Gemma-2-9B, Qwen2.5-14B) across backdoors, jailbreaks, and prompt injection (56 backdoor combinations, 3 jailbreak techniques, and BIPIA email + code-QA), LCF reduces mean backdoor attack success rate (ASR) below 1% on Qwen2.5-7B and Gemma-2 and to 1.3% on Qwen2.5-14B, detects 92-100% of DAN jailbreaks (62-100% for GCG and softer role-play), and flags 100% of text-payload injections across all eight (model, domain) cells, at 12-16% backdoor FPR and <0.1% inference overhead. A single aggregation score covers all three threat families without threat-specific tuning, positioning LCF as a general-purpose runtime safety layer for cloud-served and on-device LLMs.

The rapid integration of Large Language Models (LLMs) into Multi-Agent Systems (MAS) has significantly enhanced their collaborative problem-solving capabilities, but it has also expanded their attack surfaces, exposing them to vulnerabilities such as prompt infection and compromised inter-agent communication. While emerging graph-based anomaly detection methods show promise in protecting these networks, the field currently lacks a standardized, reproducible environment to train these models and evaluate their efficacy. To address this gap, we introduce Gammaf (Graph-based Anomaly Monitoring for LLM Multi-Agent systems Framework), an open-source benchmarking platform. Gammaf is not a novel defense mechanism itself, but rather a comprehensive evaluation architecture designed to generate synthetic multi-agent interaction datasets and benchmark the performance of existing and future defense models. The proposed framework operates through two interdependent pipelines: a Training Data Generation stage, which simulates debates across varied network topologies to capture interactions as robust attributed graphs, and a Defense System Benchmarking stage, which actively evaluates defense models by dynamically isolating flagged adversarial nodes during live inference rounds. Through rigorous evaluation using established defense baselines (XG-Guard and BlindGuard) across multiple knowledge tasks (such as MMLU-Pro and GSM8K), we demonstrate Gammaf's high utility, topological scalability, and execution efficiency. Furthermore, our experimental results reveal that equipping an LLM-MAS with effective attack remediation not only recovers system integrity but also substantially reduces overall operational costs by facilitating early consensus and cutting off the extensive token generation typical of adversarial agents.

Fine-tuning unlocks large language models (LLMs) for specialized applications, but its high computational cost often puts it out of reach for resource-constrained organizations. While cloud platforms could provide the needed resources, data privacy concerns make sharing sensitive information with third parties risky. A promising solution is split learning for LLM fine-tuning, which divides the model between clients and a server, allowing collaborative and secure training through exchanged intermediate data, thus enabling resource-constrained participants to adapt LLMs safely. % In light of this, a growing body of literature has emerged to advance this paradigm, introducing varied model methods, system optimizations, and privacy defense-attack techniques for split learning. To bring clarity and direction to the field, a comprehensive survey is needed to classify, compare, and critique these diverse approaches. This paper fills the gap by presenting the first extensive survey dedicated to split learning for LLM fine-tuning. We propose a unified, fine-grained training pipeline to pinpoint key operational components and conduct a systematic review of state-of-the-art work across three core dimensions: model-level optimization, system-level efficiency, and privacy preservation. Through this structured taxonomy, we establish a foundation for advancing scalable, robust, and secure collaborative LLM adaptation.

Public warning systems (PWS) in cellular networks enable authorities to broadcast emergency alerts to all mobile phones in a geographic region in the event of threats such as earthquakes or severe weather. If an attacker can imitate these alerts and transmit a forged warning containing fake news or phishing links, the impact could range from public panic to user compromise. In this work, we present the first open-source 5G emergency alert spoofing attack, implemented by modifying the openairinterface (OAI) radio access network (RAN) code and executed using a software-defined radio, complemented by a custom network management system to automate network and warning configuration. We conduct a detailed analysis of how different smartphones behave under various conditions. Our findings show that while devices readily display spoofed alerts, the alerting mechanism enables multiple practical attack scenarios beyond simple warning display. Finally, to address this threat, we propose and implement a lightweight cross-cell verification mechanism in OAI, in which the device compares the received warning with neighboring cell broadcasts to flag single-source alerts as suspicious.

Accurate vulnerability-inducing commit identification serves as a foundation for a series of software security tasks, such as vulnerability detection and affected version analysis. A straightforward solution is the SZZ algorithm, which traces back through the code history to identify the earliest commit that modify the vulnerable code. Unfortunately, neither the customized V-SZZ nor state-of-the-art LLM4SZZ perform satisfactorily due to the incorrect anchor selection and inadequate backtracking capability, making them far beyond a reliable usage in practice. To overcome these challenges, we propose a multi-agentic SZZ algorithm, named MAS-SZZ, that facilitates the identification of vulnerability-inducing commits through collaboration among agents. Specifically, given a CVE description and its corresponding fixing commit, MAS-SZZ summarizes the root cause of the vulnerability and employs a structured step-forward prompting strategy to localize vulnerability-related statements based on the change intent of each patch hunk. These vulnerable statements serve as anchors from which MAS-SZZ autonomously traces backward through the repository's history to find the commit that first introduced the vulnerability. Extensive experiments show that MAS-SZZ outperforms the state-of-the-art baselines across datasets and programming languages, achieving F1-score gains of up to 65.22% over the best-performing SZZ algorithm.

A t-private n-server Information-Theoretic Distributed Point Function ((t,n)-ITDPF) allows one to convert any point function f_{alpha,beta}(x): [N] -> G into n shares (secret keys), such that each server can compute an additive share of f_{alpha,beta}(x) with a key while any <= t servers learn absolutely no information about the function. This paper constructs a novel share conversion based on the private information retrieval (PIR) of Ghasemi, Kopparty, and Sudan (STOC 2025) and proposes a perfectly secure 1-private ITDPF with output group G = Z_p, where p can be any prime. Compared with the existing perfectly secure ITDPFs for the same output group, the proposed ITDPF is more efficient with asymptotically shorter secret keys.

Fast Adversarial Training (FAT) has attracted significant attention due to its efficiency in enhancing neural network robustness against adversarial attacks. However, FAT is prone to catastrophic overfitting (CO), wherein models overfit to the specific attack used during training and fail to generalize to others. While existing methods introduce diverse hypotheses and propose various strategies to mitigate CO, a systematic and intuitive explanation of CO remains absent. In this work, we innovatively interpret CO through the lens of backdoor. Through validations on pathway division, diverse feature predictions, and universal class distinguishable triggers in CO, we conceptualize CO as a weak trigger variant of unlearnable tasks, unifying CO, backdoor attacks, and unlearnable tasks under a common theoretical framework. Guided by this, we leverage several backdoor inspired strategies to mitigate CO: (i) Recalibrate CO affected model parameters using vanilla fine tuning, linear probing, or reinitialization-based techniques; (ii) Introduce a weight outlier suppression constraint to regulate abnormal deviations in model weights. Extensive experiments support our interpretation of CO and show the efficacy of the proposed mitigation strategies.

Cross-chain bridges, the critical infrastructure of the multi-chain ecosystem, have become a primary target for attackers, resulting in over $2.8 billion in losses due to subtle implementation flaws. Existing defenses, such as bytecode-level static analysis, are ill-equipped to handle the semantic complexity of cross-chain interactions, while LLM-based approaches, which can understand source code, struggle with hallucinatory reasoning over complex, multi-contract dependencies. In this paper, we propose GoAT-X, a framework that shifts automated cross-chain smart contract codebases auditing from heuristic pattern matching toward systematic first-principles verification. GoAT-X structures the audit process as a Graph of Auditing Thoughts, explicitly mirroring how human experts decompose, reason about, and validate security logic. By anchoring LLM reasoning in statically extracted data flows and explicitly linking abstract security properties to concrete code implementations, the framework constrains semantic reasoning within well-defined structural and state boundaries. Within this constrained space, GoAT-X treats missing constraints and adversarial bypass paths in cross-chain logic as first-class vulnerability targets and dynamically explores reasoning paths to identify exploitable semantic gaps. We evaluate GoAT-X on a comprehensive benchmark covering all known cross-chain token transaction attacks. GoAT-X achieves 92% recall on fine-grained audit points and 95% coverage of vulnerable projects, while identifying 117 confirmed risks in the wild with low operational cost, establishing a new standard for scalable, logic-driven cross-chain security.

Fast Adversarial Training (FAT) has proven effective in enhancing model robustness by encouraging networks to learn perturbation-invariant representations. However, FAT often suffers from catastrophic overfitting (CO), where the model overfits to the training attack and fails to generalize to unseen ones. Moreover, robustness oriented optimization typically leads to notable performance degradation on clean inputs, and such degradation becomes increasingly severe as the perturbation budget grows. In this work, we conduct a comprehensive analysis of how guidance strength affects model performance by modulating perturbation and supervision levels across distinct confidence groups. The findings reveal that low confidence samples are the primary contributors to CO and the robustness accuracy trade off. Building on this insight, we propose a Distribution-aware Dynamic Guidance (DDG) strategy that dynamically adjusts both the perturbation budget and supervision signal. Specifically, DDG scales the perturbation magnitude according to the sample confidence at the ground truth class, thereby guiding samples toward consistent decision boundaries while mitigating the influence of learning spurious correlations. Simultaneously, it dynamically adjusts the supervision signal based on the prediction state of each sample, preventing overemphasis on incorrect signals. To alleviate potential gradient instability arising from dynamic guidance, we further design a weighted regularization constraint. Extensive experiments on standard benchmarks demonstrate that DDG effectively alleviates both CO and the robustness accuracy trade off.