Papers

Cybersecurity combines high-stakes analysis with complex technical language, making it an impactful and challenging domain for LLMs. We present MiST (Mid-trained Security Transformer), a suite of 8B and 32B models that achieve strong performance on public cybersecurity benchmarks. We use mid-training as an intermediate adaptation stage between general pre-training and cybersecurity training. Rather than performing continual pre-training over large volumes of raw domain text, we curate a compact, expert-vetted seed corpus, and transform it into high-quality domain-specific synthetic training data. The final MiST checkpoints improve mean cybersecurity accuracy by +13.1 and +8.6 absolute percentage points over the corresponding Qwen baselines for 8B and 32B, respectively, corresponding to relative gains of +27.0% and +15.8%. Ablation results further show that these cybersecurity gains arise in the mid-training and supervised fine-tuning stages through a combination of the synthetic data generation flows. Furthermore, we show that MiST provides a stronger initialization for downstream task-specific fine-tuning adaptation and reinforcement learning.

Cryptographic dependence predates the quantum era, but the migration to post-quantum cryptography (PQC) opens a rare window to reshape it, because the algorithms, implementations, hardware, and standards adopted now can lock in dependence or sovereignty for decades. This paper introduces the Readiness-Sovereignty Capability Model (RSCM), a national measurement model that operationalizes PQC readiness together with cryptographic sovereignty, which current maturity models score only as readiness and the sovereignty literature defines without measuring. RSCM decomposes sovereignty into three distinct constructs, indigenous cryptographic capacity, indigenous post-quantum control, and external dependency, and certifies a post-quantum maker only through a gate requiring demonstrated, institutionally sustained creation in at least one core layer, whether design, implementation, or validation. Applying it to fifty-seven documented cryptographic actors coded from cited public evidence, and testing that coding with an independent second coder, a plausible-state bootstrap, and convergent-validity checks, we find that twenty countries clear the gate, fifteen as full-stack makers and five as research makers, eleven hold strong general capacity without post-quantum control, one is a ready adopter, and twenty-five are dependent. The gate cells show substantial weighted agreement, a quadratic-weighted kappa of 0.71, and the maker classification is stable in its core though uncertain at the threshold. Readiness tracks independent cyber indices at rank correlations up to 0.70, while post-quantum creation shows no significant correlation with the commitment index, a rank correlation of only 0.22 that separates control from readiness. The paper contributes the framework, the evidence-graded assessment, and policy directions for building indigenous quantum-safe capacity.

How does a multi-agent system evolve from a local deviation into collective loss of control? We propose an epidemic explanation organized around accidental mutation, contagion, and recovery. A spontaneous deviation creates a seed; communication enables other agents to adopt and retransmit its unsafe strategy; collective failure can emerge when propagation outpaces correction and containment. Thus, rare individual deviations can coexist with substantial collective risk. Motivated by reported OpenAI agent coordination incidents, we examine two ingredients of this mechanism. A deployment audit identifies implicit communication paths between nominally independent evaluation runs and verifies transport through a default Docker backend. RogueHandoff-20, a benchmark of 20 executable scenarios, tests recipient susceptibility by injecting unsafe trajectories generated by a modified Qwen-27B route. Across four native-pending routes, executed harm is 0-5% on normal tasks and 40-95% after injection, exceeding paired direct malicious requests by 5-45 percentage points. These results support low observed baseline harm alongside high conditional susceptibility; they do not establish natural rare-event rates or demonstrate an autonomous cascade. The account motivates complementary defenses: strengthen resistance and recovery alongside prevention of spontaneous deviations, and audit and restrict unintended communication paths that can turn local failures into collective loss of control.

Encrypted communication protects sensitive user data but can facilitate harmful or unlawful exchanges, creating a trade-off between detecting dangerous messages and preserving end-user privacy. To address this, we propose SEEK, a practical and efficient encrypted keyword-search protocol for privacy-preserving messaging that combines homomorphic encryption with secure two-party computation (2PC). SEEK first partitions messages into ciphertext fragments with the minimum sufficient overlap, then homomorphically correlates them using encrypted keyword trapdoors. For long messages, this design can reduce sender-side encryption and upload overhead by up to two orders of magnitude over state-of-the-art baselines. It supports ASCII case-insensitive matching with one fixed-size encrypted trapdoor and one homomorphic multiplication per fragment, yielding up to 5.47x faster correlation computation than the strongest fragmentation-based baselines. SEEK then invokes 2PC-based selected decoding, blinded zero testing, and secure aggregation, revealing only the keyword presence-or-absence bit while hiding the keyword, its length, message contents, match counts, and locations. SEEK achieves 100% accuracy under case variations that result in exact-matching failures, without requiring additional trapdoors or online communication. We further realize SEEK as an end-to-end web and cross-platform mobile application. Prototype evaluation on a weekly messaging history yields an online computation time of 1.92 s per search, demonstrating the practical feasibility and efficiency of SEEK.

Modern fuzzers use code coverage as feedback to guide their exploration which has proven to be an effective strategy for driving exploration. However, this strategy overlooks inputs that may be interesting to the target program even without uncovering new code paths. Fortunately, prior research has shown that annotations generated by human domain experts can provide additional feedback, guiding the fuzzer towards interesting parts of the program. In this paper, we replicate experiments presented in IJON and extend them to real-world vulnerability detection at scale. To mitigate the scalability challenge, imposed by the need for human domain expertise, we propose utilizing LLMs to automatically generate annotations. We demonstrate the applicability of LLMs for this purpose and observe that LLMs can generate annotations that perform comparably to human-generated annotations. Motivated by this finding, we design AIJON, a system that leverages LLMs to automatically generate IJON-style annotations. We evaluate AIJON on the Magma benchmark and surprisingly observe that annotation-based fuzzing does not perform strictly better than AFL++. We conduct several experiments to identify the cause of our results and identify key insights regarding the impact of annotations on fuzzing campaigns, including their effect on the energy distribution of the fuzzer. Notably, we observe that LLMs can generate annotations that achieve comparable results to human generated ones, thus opening the door for future research to perform further studies on the impact of annotations at scale.

Agentic browsers can execute security-sensitive actions under a user's authenticated session, making indirect prompt injection and deceptive confirmation interfaces a direct threat to action integrity. Existing human-in-the-loop (HITL) safeguards are insufficient when the approval prompt itself can be influenced by untrusted page content or model-generated text. We present the Verifiable Action Card (VAC), an architectural defence that reconstructs approval information from the ground-truth pending browser action and trusted intent provenance, renders it out-of-band in the trusted browser chrome, and binds approval to the exact action re-verified at dispatch. VAC combines provenance fencing, a ground-truth action descriptor, default-deny confirmation, provenance-aware risk gating, and execution binding. We implement VAC in a complete agentic browser and evaluate it on a 24-scenario benchmark covering confused-deputy attacks, Lies-in-the-Loop dialog forging, indirect prompt injection, adaptive action substitution, provenance evasion, and legitimate tasks. Across the evaluated LLMs, attack success without VAC ranges from $68\%$ to $100\%$, whereas VAC reduces attack success to $0\%$ on every model, with $78\%$ legitimate-task completion and a $0\%$ false-block rate. These results show that grounding approval in the action that will actually execute provides architectural protection against security failures that prompt-level defences and conventional HITL confirmation cannot reliably prevent.

Blockchain-enabled Internet of Things (IoT) systems integrate smart contracts with embedded devices to support decentralized device management and access control. Their security therefore depends jointly on the logic of on-chain contracts and off-chain device firmware. Logic flaws in either layer can violate the same system invariants, such as unauthorized access, improper state changes, or unguarded privileged operations. Existing approaches rely on contract analysis, firmware analysis, and graph-based vulnerability detection. However, these methods typically focus on a single layer or artifact and often depend on predefined vulnerability patterns, emulation fidelity, or homogeneous representations that obscure security-relevant component roles. They also lack a unified architecture that supports different security tasks while remaining deployable on resource-constrained gateways. To address these limitations, we extend MA-HGAT into a cross-layer multi-agent heterogeneous graph attention framework that models contracts, firmware artifacts, device fleets, and transaction streams with a unified four-role, nine-relation schema. Role-aligned agents exchange heterogeneous evidence through cross-attention, while graph-, link-, and node-level heads support multiple detection tasks and a role-based gateway–cloud partition enables lightweight edge inference. MA-HGAT thus provides a unified and deployable framework for detecting logic vulnerabilities across the contract and device layers of blockchain-enabled IoT systems.

Adaptive security at the network edge increasingly relies on automated planners, including rule-based controllers, learned policies, and LLM-assisted agents, that translate observations into enforcement actions. Once such a planner can influence live policy state, syntactic validity is not enough. A semantically wrong action, produced from incomplete or manipulated observations, can be faithfully executed by an enforcement substrate that cannot judge mission context. We address this problem by treating the boundary between planner output and kernel enforcement input as the primary security object. We propose a split-control architecture in which an untrusted planner emits typed security intents, a deterministic governor checks each intent against safety, resource, temporal-stability, and proportionality invariants, and only admitted actions are bound to signed receipts and compiled into pre-installed eBPF map updates. The paper formalizes this trust-boundary problem, defines three threat classes, develops the governor admission predicate, and reports an end-to-end prototype. Across rule-based and LLM-assisted planners on a Raspberry Pi 5 testbed connected to the university 5G Test Network, the governor admits, rejects, and bounds intents at microsecond cost without disrupting protected-flow regularity. The contribution is conceptual as much as empirical: adaptive security does not need to trust the author of an action. It needs a mediation boundary that decides whether the action is admissible.

The availability and consistency of online services remain vulnerable due to Distributed Denial of Service (DDoS) attacks. These attacks are evolving by adopting more complex strategies to evade traditional network security systems. Despite the effectiveness of machine learning models in detecting DDoS traffic, targeted adversarial attacks can degrade their classification accuracy. This work proposes a robust detection framework that integrates generative adversarial modelling with advanced machine learning models. We trained Random Forests, Deep Neural Ensembles, and Transformer-based models using the CICDDoS2019 dataset to establish the frameworks baseline performance. To enhance the models defensive capacity, we generated synthetic adversarial flows that simulate potential evasion attempts and adversarial traffic using a Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP). Then, we combined the generated traffic with benign and malicious traffic to construct hybrid datasets to train the models to learn more generalizable decision boundaries. The experimental results indicate that the proposed methodology significantly enhances detection accuracy and resilience, especially against unseen adversarial traffic. We also tested the designed framework using real-world generated traffic, which demonstrates its capability in practical settings. The scalable and efficient solution against adversarial DDoS attacks, introduced in this work, paves the way towards more resilient and adaptive network defense systems that combine generative adversarial augmentation with recent advances in learning models.

For every constant $c > 0$, we unconditionally construct witness encryption for NP in the classical generic cyclic group model with the following properties. For SAT instances of size $n$, encryption runs in time poly$(n)$, and any satisfying assignment can be used to decrypt in poly$(n)$ time with correctness error $2^{-n}$. If no satisfying assignment exists, every generic adversary making at most $n^{(\log n)^c}$ group queries has distinguishing advantage at most $n^{-(\log n)^c}$. Along the way, we give the first superconstant-factor NP-hardness of approximation result for homogeneous MinRank under randomized polynomial-time reductions. We obtain a gap of $(\log n)^c$ for any prescribed constant $c>0$, even when YES instances are required to have a rank-one matrix of the form $\mathbf u\mathbf v^T$ where $\mathbf v$ is a 0/1 vector. The main results of this paper were discovered by Codex / ChatGPT 6 Astra Ultra using harness components from the UCLA Moonshot Harness Project, in response to a general strategy for witness encryption suggested by the authors. See the end of the introduction for a discussion.

Agentic AI systems plan, invoke tools and act with limited supervision; they are now both the subject of audits and, increasingly, the auditor. Independence, the foundation of assurance,is still applied to them as a binary. We argue that it must be graded along three orthogonal axes: principal independence (who controls the auditor), substrate independence (an auditor sharing the auditee's foundation-model family, toolchain or guardrails fails with it) and evidence independence (whether evidence is attestable rather than self-reported). Each axis has precedent; the contribution is to grade all three on a single audit, aggregate them by the weakest link, and apply the same rubric when the auditor is itself an agent. We give the model a formal basis by transplanting the beta-factor model of common-cause failure from reliability engineering, a seven-step protocol whose outputs a third party can verify, a structural detectability analysis of a procurement-controls agent audited at three grades, and a Monte Carlo study of the model in which a conventional internal audit of an agent-a real audit team, a second agent, provider logsp-surfaces 5.9% of the faults it could in principle see and none at all in half the fault classes. We map the triple to the EU AI Act as amended, ISO/IEC 42006, UK public-sector risk-management guidance and audit-regulator practice.

The Model Context Protocol (MCP) enables LLMs to invoke external tools, but every tool interaction exposes the model to attacker-controlled text through multiple input channels (tool descriptions, tool results, sampling messages) that share a single context window without privilege separation. In this paper, we present a framework to measure the trust profile of an arbitrary LLM based on a variety of payload framings sent through different channels. Following this assessment, we devise cross-channel fragmentation attacks that distribute seemingly benign payloads across two or three channels; no individual channel carries a complete injection, yet the LLM compiles the fragments into credential exfiltration. We evaluated our attacks across 12 frontier models, three production clients, and six payloads, totalling over 15,000 trials. Our evaluation reveals that cross-channel attacks are an unexplored attack surface: models that fully resist single-channel injection (0% compliance) exfiltrate sensitive data at up to 100% under two-channel fragmentation (e.g., GPT-4o, Llama 70B, Composer 2, Haiku 4.5). We further demonstrate value-aligned exploitation, where a tool's stated purpose requires the data the attacker targets, and a sampling system prompt override that injects persistent instructions via VS Code's MCP implementation. Finally, we evaluated our attacks against seven third-party MCP security tools and three prompt-based defenses. All tools failed to detect fragmented payloads, and prompt defenses proved model-specific rather than universal.

Operational Technology (OT) environments require realistic, reproducible security datasets, yet existing approaches often lack automation, multi-source data capture, and sufficient documentation for reuse. This paper presents ISIA-AF, a modular attack framework for orchestrating reproducible attack execution and automated dataset generation on industrial systems. The framework coordinates distributed attack clients, records network traffic and operational data, ultimately leading to a multi-source dataset. We derive functional and non-functional requirements from prior work and stakeholder discussions, and realise the framework following a design science research approach. A case study on the ISIA testbed, comprising a real industrial system and a simulated process, demonstrates how the framework supports centralised control, low communication overhead, and flexible deployment across network segments. The result is a practical basis for generating extensible, multi-source OT security datasets for intrusion detection research.

Cross-chain bridges enable interoperability, but they also break the transaction trails needed to trace illicit funds. Third-party investigators typically cannot access the source-to-destination mappings maintained by bridge backends, and our survey of 131 bridges finds that only 16.79% provide complete public tracking. Existing approaches depend on official APIs, EVM-specific assumptions, or fragile temporal heuristics, limiting their ability to trace transfers across heterogeneous ledgers. We present XSplicer, an evidence-driven system for reconstructing cross-chain transaction correspondence (xTCR) without privileged access to bridge backends. XSplicer derives unified semantic specifications from public protocol documentation and transaction examples, translates them into lightweight parsers and verifiers, and links source and destination transactions by prioritizing hard evidence and using soft clues only when necessary. We evaluate XSplicer on seven bridge protocols spanning EVM, Bitcoin, and Solana. XSplicer achieves 92.5% global recovery rate and up to 98.61% on individual protocols. Under adversarial noise, its hard-evidence verifier retains the correct match in 100% of tested cases, while soft-clue matching degrades as ambiguity increases. In two real-world case studies, XSplicer recovers more than 1,900 historical transaction pairs after Multichain ceased operations and identifies 754 illicit cross-chain transfers worth 105.6 million USD in the Bybit laundering incident. These results show that public protocol invariants can support practical cross-chain forensics without privileged bridge mappings.

Learning-enabled robotic manipulation increasingly relies on robot simulators for policy training and evaluation before real-world deployment. Inside a simulator, a 3D asset contains two separate geometries: a visual mesh used for rendering and a collision mesh used for physical interaction. For computational efficiency, the collision mesh is deliberately a coarse approximation that need not have the same geometry as the visual mesh, a legitimate and pervasive discrepancy we call the Visual–Collision Gap (V–C Gap). We show that the V–C Gap opens a new and practical attack surface, and propose Collision Mesh Poisoning (CMP), the first poisoning attack against robotic manipulation delivered through the 3D asset supply chain. An attacker modifies only the collision mesh of a 3D asset, leaving the visual mesh and all other components unchanged. A policy trained and evaluated with the poisoned asset behaves normally throughout simulation, yet degrades, fails, or creates physical safety risks once deployed in the real world. Since current asset review practices cover malware, copyright, and format compliance, but not visual–collision consistency, poisoned assets can be distributed through legitimate supply chain channels. We evaluate several defenses and our results show that they are insufficient to defend against CMP, highlighting the need for new defenses.

AI-driven penetration testing has been demonstrated with premium frontier models such as GPT-4, but the per-engagement token cost makes continuous, automated testing unaffordable for the smaller organisations that need it most. This paper presents PentestChain, a ten-phase automated penetration testing framework that couples a curated, deterministic exploit map with a cost-aware AI cascade-a local Ollama model (qwen2.5-7b) first, then free-tier OpenRouter and Cerebras, with a rule-based fallback that always produces output-and exposes the full pipeline through a Model Context Protocol (MCP) server with eleven tools. We make three contributions. First, we treat US-dollar cost per engagement as a measured, first-class evaluation metric and show that a 7B-parameter local model, kept off the critical path by a deterministic backbone, sustains end-to-end operation at zero measured paid-API cost. Second, we analyse the attack surface that an MCP-exposed offensive engine introduces, grounding a four-position threat model in the 2025 MCP incident record (the CVE-2025-6514 remote-code-execution flaw in mcp-remote, the postmark-mcp supply-chain backdoor, and the tool-poisoning-rug-pull-line-jumping class), and contribute four mitigations. Third, we specify a reproducible, containerised evalua-tion protocol aligned with the standardised testbeds now expected at top-tier venues-AutoPenBench, a Cybench subset, and the PentestGPT 182-sub-task benchmark-with multi-trial statistics (more than 10 trials per configuration, pass-at-k, non-parametric significance tests and effect sizes) and direct, same testbed reproduction of the PentestGPT and PentestAgent baselines rather than citation of their published numbers. On the legacy targets measured to date, the framework detected 26 services, enriched 34 CVEs, produced

Graph anomaly detection aims to identify graph structures (e.g., nodes, edges, or subgraphs) that deviate significantly from expected patterns, which supports critical applications in fraud detection, spam identification, network intrusion, etc. Despite the growing methods in the field, existing approaches follow a one-model-per-dataset paradigm, limiting their transferability across diverse real-world scenarios due to task heterogeneity, label scarcity, and domain variability. In this work, we introduce FoundAna, a GNN-assisted Foundation Model for Graph Anomaly Detection - the first foundation model framework designated for generalizable, cross-graph anomaly detection by combining GNNs and transformers. FoundAna integrates an anomaly detection-specific GNN component with a standard transformer encoder augmented by four complementary positional encodings, which enable the model to capture both local and global structural information. Specifically, the positional encoding enriched node representations are passed through attribute and adjacency decoders, and the reconstruction errors serve as the anomaly score. Extensive experiments on nine benchmark datasets spanning financial, social, and citation network domains demonstrate that FoundAna consistently outperforms state-of-the-art baselines. The code implementation and Supplementary materials are here: https://github.com/FoundAna331/FoundAna.

A decomposable randomized encoding (DRE) for a function $f$ allows $n$ parties, using shared randomness, to encode their individual inputs locally so that the collection of encodings reveals $f(x_1,\ldots,x_n)$ and nothing else. DREs are widely used in efficient multiparty computation. Their main complexity measure is size, the total bit length of the local encodings. Yet the optimal DRE size remains poorly understood even for the $n$-bit OR function. We prove the first superlinear lower bound for OR and, more generally, for every non-periodic symmetric function. Under an additional symmetry assumption, we prove a sharp $\Omega(n\log n)$ lower bound for OR, matching the classic construction of Feige, Kilian, and Naor (STOC 1994). We also prove the first $\Omega(n^2)$ lower bound on DRE size for non-explicit Boolean functions.

Natural-language datasets support many downstream applications and research studies, but releasing text can reveal sensitive global properties of the underlying data source, such as the proportion of records associated with a particular gender, diagnosis, or political stance. Existing work has largely focused on property inference attacks that recover such global properties, while defenses for protecting these dataset-level secrets remain limited. Differential privacy, although effective for protecting individual records, provides only weak protection for aggregate properties. We propose Randomized Quantization for Text (QuanText), a training-free and large-language-model-agnostic data release mechanism that protects global secrets in textual datasets while preserving data utility. Given a dataset-level secret, such as the proportion of records with a particular diagnosis, and attributes whose utility should be preserved, such as topic and sentiment, QuanText perturbs both the secret distribution and the distributions of correlated attributes. It does so by constructing candidate release distributions over secret and non-secret attributes, randomly selecting a candidate sufficiently close to the private empirical distribution, and rewriting each private text sample to match the selected distribution using attribute-related snippets from the original text. QuanText is inspired by the Statistic Maximal Leakage (SML) framework, which bounds leakage about a secret function of a data distribution. Under idealized conditions, we show that QuanText satisfies an SML guarantee. Since these conditions may not hold exactly in practice, we also evaluate QuanText empirically on real-world datasets. Our results show that QuanText achieves a better empirical privacy-utility trade-off than competing data generation baselines.

Recent work suggests that obscure language registers can weaken large language model refusal behavior, especially when paired with black-box prompt optimization. It remains unclear whether failures stem from non-standard surface form, culturally grounded framing, or optimization over an expressive prompt-strategy space. We study Chinese registers by extending a classical-Chinese red-teaming framework to Shanghainese and Cantonese and running a 36-cell ablation across surface forms, strategy-bank variants, and two target models. The main finding is corrective: dialectal surface form is neither necessary nor sufficient for high attack success. Non-optimized English, Mandarin, and naive dialect translations remain below 8% attack success rate, whereas all conditions that retain an optimizer-controlled strategy bank reach 98–100%. A culture-neutral generic strategy bank reaches the same ceiling at near-single-query cost, further indicating that strategy-bank expressiveness, rather than dialectal or cultural content, accounts for most of the observed effect. Dialect choice still affects query efficiency and response severity, and qualitative coding identifies recurring high-level mechanisms such as semantic glossing and procedural scaffolding. We qualify the absolute ceiling-level attack success rate by showing sensitivity to judge calibration and noting that the design does not fully separate one-shot strategy construction from iterative search.

Mixers protect privacy by concealing deposit–withdrawal links, but are also abused to launder illicit funds. Existing anti-money laundering studies do not specifically target mixer laundering, while mixer research focuses on deanonymization rather than identifying laundering-related transactions. Public reports remain fragmented, leaving no public case-level dataset for systematic measurement and detection. To fill this void, this paper presents the first comprehensive study of mixer laundering on Ethereum. We first construct MixLaunder, the first public case-level dataset of mixer laundering. It covers 27 cases involving Tornado Cash and Railgun from 2020 to 2025 and labels 9,300 laundering-related transactions with case identities and observable upstream and downstream fund flows, including deposits totaling approximately \$1.1 billion. By comparing these transactions with background mixer usage, we identify five common strategies, showing that laundering evidence spans complementary behavioral and fund-flow contexts, while same-case activity is locally tight but weakly connected across bursts. Our analysis further reveals coverage gaps in mixer-side risk screening and representative deanonymization heuristics. Guided by these findings, we develop MixGuard, which combines tri-view representation learning with two-stage grouping for transaction-level detection and case-aware grouping. Under strict case-level holdout evaluation, MixGuard outperforms representative baselines, achieving 97.89% detection precision and 98.73% group purity, while its top ten groups cover 95.09% of each case's transactions on average.

Hardware Trojans are malicious modifications that compromise functionality or leak sensitive data. They pose a severe threat, particularly when inserted at the Register Transfer Level (RTL). After synthesis, these Trojans are often concealed by optimizations in gate-level netlists. Recent efforts, including the ICCAD 2025 contest, emphasize golden-chip-free detection using machine learning (ML) on labeled netlists. In this work, we show that RTL Trojans exhibit stable structural and signal-flow patterns post-synthesis, enabling effective detection through targeted heuristics rather than generic ML feature learning. We propose LoRD, a lightweight heuristic-based approach that exploits these distinctive subgraph signatures, achieving near-perfect detection and localization on the contest testcases. Com- pared to a transformer-based ML baseline and top five teams, LoRD achieves on-average a score of 2.957 (out of 3) for Trojan- implanted designs without the data and tuning overhead.

Blockchain based prediction markets combine offchain order management with onchain settlement. This architecture supports user controlled custody, since users keep funds in wallets or smart contracts while submitting signed orders to an offchain order book. However, it creates an atomicity gap. An order may be valid when accepted or matched offchain, but become invalid before the corresponding onchain settlement transaction is executed. This behavior, often called ghost filled orders by the community, can cause trades that appear filled offchain to fail onchain. This paper studies this atomicity gap through a case study of Polymarket. We show how the delay between offchain order acceptance and onchain settlement allows adversaries to invalidate unfavorable orders after observing market outcomes or price movements. We then quantify the scale and financial impact of this behavior over a nine-month period from August 12, 2025, to May 22, 2026, using 1.8 million reverted transactions involving Polymarket official smart contracts. Our analysis separates attacker profit from market and user loss, and develops conservative measurement rules to avoid overclaiming impact. We further develop a methodology for testing other blockchain based prediction markets and apply it to three additional markets. We find that all three are vulnerable to the same class of attack, and that one design enables direct attacker profit. We have reported the findings to all three projects; one project had acknowledged the issue at the time of the study.

We study cross-chain arbitrage when autonomous AI agents, rather than humans or bots, are the searchers. We model agents as both arbitrage extractors and Maximal Extractable Value targets, derive the optimal trade size for a risk-averse agent under mean-variance utility with stochastic bridge delays, and formalize multi-chain path selection as a belief-weighted online learning problem whose belief estimates converge under a Robbins-Monro schedule. Using 23,000 Uniswap V3 swap events across Ethereum, Arbitrum, and Base, we find that Ethereum-Arbitrum price gaps average 0.044% at 10-second resolution and Arbitrum–Base gaps average 0.013%, so $10,000 trades clear in 63% of L2-L2 windows via CCTP while L1-L2 routes require $50,000 or more for comparable viability. Our adaptive path-selection algorithm outperforms standard baselines by 11% on average, and moderate randomization cuts MEV exposure by over 50% with only modest profit loss.

Heterogeneous cooperative perception (CP) enables connected vehicles with diverse sensor setups to share spatial awareness via compact feature maps, where receivers reconcile these maps using learned translation modules for fusion and inference. Prior attacks against CP in a homogeneous setting reveal that the data exchange introduces a critical attack surface: a single malicious agent can transmit crafted features that erase real objects from a neighbor's fused scene. Yet, it is widely hypothesized that heterogeneity naturally defends against these attacks, as the attacker lacks knowledge of the victim's detector and the translation module scrambles adversarial gradients. We demonstrate that this protection is largely an illusion. Using a matched-objective harness to standardize the perturbation budget, objective, and forward path, we show that properly tuned iterative attacks close or reverse the apparent robustness gap. However, these optimization-based attacks require ground-truth labels and iterative backpropagation, meaning they do not represent a practical field threat running in real-time. To bridge this gap, we introduce HetPoison, a learned generator that crafts a removal perturbation in a single, label-free forward pass. HetPoison transfers across major heterogeneous designs without requiring access to the victim's detector, matching or exceeding the effectiveness of expensive optimizer-based attacks. Since heterogeneity itself is not a defense, we propose HetShield, a lightweight trust layer that validates the spatiotemporal consistency across features, recovering 83–95% of the accuracy degraded by attacks, outperforming prior art.

Users increasingly turn to Large Language Models to answer a variety of questions, including cybersecurity questions. We study how personalization strategies can help improve the effectiveness of answers to questions asked to an LLM-based cybersecurity assistant. Beyond accuracy, we focus on the understandability, actionability and, most importantly, motivating power of answers, given how often users fail to follow cybersecurity recommendations. Specifically, we investigate four personalization strategies, ranging from static user profiles to interaction-history-based personalization, using a corpus of 1,045 real-world cybersecurity questions and a 7-day deployment involving 57 participants and 1,066 user questions. Across both a large-scale automated LLM-based evaluation and human evaluation, conversation-based personalization is consistently favored in comparative ratings of perceived helpfulness and likelihood of following security advice. Importantly, the relative trends observed in the LLM-based evaluation align with those obtained from human evaluation, suggesting that LLM-based evaluation can provide a scalable mechanism for comparing personalization strategies before costly user studies. These results indicate that behavior-driven personalization is a promising direction for LLM-powered cybersecurity assistants and highlight the value of combining LLM-based and human evaluation when studying personalized language-model systems.

Thompson's "Reflections on Trusting Trust" showed that a compiler can be poisoned to reinsert its own backdoor, so that even recompiling clean source reproduces the Trojan. Today, substantial coding work is done by AI coding agents – and increasingly, those agents generate new versions of themselves. We reconsider Thompson's attack when the "compiler" is a self-modifying coding agent. Can an adversary supply poisoned benchmarks to the agent's self-evaluation and self-improvement process to induce future versions of the agent to write vulnerable code on clean, held-out tasks? We instantiate this attack against three recently proposed self-modifying coding agents: the Darwin Gödel Machine (with our experimental modifications), the Self-Improving Coding Agent, and Hyperagents (both substantively unmodified). We demonstrate successful proofs-of-concept: for example, with Hyperagents powered by Sonnet 4.5, our poisoned benchmark leads the agent to self-evolve instructions that disable HTTPS certificate validation on neutral URL-fetching tasks. From our experiments, we distill properties of the vulnerability, benchmark, model, and agent scaffolding that are sufficient to enable a benchmark poisoning attack. Moreover, we show that contamination often persists even when a poisoned agent is subsequently evolved against clean benchmarks. We discuss defensive directions and argue that self-modifying coding agents must be designed to be more resilient to such attacks.

The Binary Tree Mechanism is a standard algorithm for differentially private continual counting, but its asymptotic optimality under pure differential privacy has remained unresolved since its introduction. We resolve this question. For fixed $0 < \varepsilon \le 1$, we prove asymptotically tight lower bounds of $\Omega(\log^2 n)$ for worst-case expected $\ell_\infty$ error and $\Omega(\log^3 n)$ for mean and maximum per-coordinate expected squared error. These bounds hold for arbitrary mechanisms, even when the entire stream is available in advance. The same lower bounds hold under approximate differential privacy whenever $\delta\le n^{-c}$, for any fixed $c>0$. Our lower bounds match the Binary Tree Mechanism instantiated with Laplace noise, establishing its asymptotic optimality under both pure differential privacy and approximate differential privacy in the standard regime of $\delta \ll1/n$. Our proof uses a single hard distribution with a bounded exponential score on a tree. A simple modification of the score allows the same framework to establish tight lower bounds for all three error measures.

Multi-agent LLM systems increasingly automate tasks involving agents with different levels of privilege, creating a security risk in which a compromised low-privilege agent can influence a higher-privilege agent and trigger an unauthorized action. We study attack propagation in a four-agent LangGraph pipeline comprising a Supervisor, Researcher, Validator, and Executor. We evaluate shared-memory poisoning and indirect prompt injection through a forged approval embedded in a retrieved document. We compare the Validator's judgment with an independent authorization layer using task-bound signed tokens and a separately verified policy oracle. Our contribution is an empirical study of attack propagation, a component-level ablation of the authorization boundary, and the Judgment Bypass Rate (JBR), which measures compromise at the attacked agent rather than at the final action. Across three seeds and 60 labeled tasks, memory poisoning reaches execution in every undefended trial. With authorization enabled, it achieves 100% JBR but 0% Unsafe Action Rate, showing that the Validator can remain compromised while execution is contained. Against an attacker possessing the signing secret, the policy oracle provides the observed containment, while an independently authored least privilege policy preserves this result. An Observer layer reduces the false-positive rate for agent hijacking from 49% to 7% without weakening execution-level security. These results show that structural authorization can contain compromised agent behavior even when upstream LLM judgment fails.

Durable-execution frameworks commonly journal execution steps of a long-running function and reconstruct state by deterministic replay; the journal is therefore the workflow's authoritative state, and in current deployments it typically resides with whichever provider hosts the run. This paper examines what changes if the journal is instead an authenticated, append-only log replicated among the parties to the workflow. We propose an architecture in which runs are chains of single-writer epoch cores; executor succession is governed by quorum-finalized closure certificates; durability, timeout, and checkpoint decisions are carried by explicit certificates rather than implicit trust; and large payloads travel a separate content-addressed distribution plane. A threat model separates what signatures establish (authorship and order) from what they do not (truth of recorded effects). We state the design as a set of typed artifacts and invariants, identify which claims are established by existing systems and which are proposals requiring validation, and define the prototype experiments, including specific adversarial cases, that would validate or falsify the proposal's central mechanisms.