CyberSec Research

APort Vault is a benchmark for payment authorization in tool-using AI agents. It replays 4,371 attacks written by humans against a live payment agent during a public capture-the-flag event, across 14 models from 8 labs, five policy configurations and two replay tracks, with and without a deterministic pre-action check implementing the Open Agent Passport (OAP) specification. 225,964 evaluations completed. We report five distinct events per evaluation, because collapsing them is how an agent benchmark produces a number that does not survive review. Requests are common and their rate differs far more across configurations than across models, though each attack exists at exactly one configuration so policy and attack cohort vary together: 10.9% of model-alone evaluations at Level 1, 3.0% at Level 2, 0.1% at Level 3, 79.4% at Level 4. On the 1,293 Level 4 prompts, each evaluated on every model, request rates run from 71.2% to 84.3%, and 809 prompts (62.6%) elicited a request from all fourteen models, each ending in a successful payment to the level's allowlisted recipient. The authorization boundary is where the conditions diverge. At Levels 2 to 4, transfers to recipients the passport did not permit number 140 of 76,842 with the model alone and 0 of 69,297 behind the layer, and 105 against 0 on 68,970 matched model, prompt and track triples. The zero spans 790 source sessions, giving a per-session upper bound of 0.38%. It was not obtained by refusing payments: 25,370 payments executed behind the layer, while the policy denied 187 of the 25,640 transfer calls it evaluated, 148 of them for a forbidden recipient. We release the 225,964 evaluations, the level passports, the scoring code and the analysis script at huggingface.co/datasets/aporthq/vault-benchmark-v1 .

Given a supersingular elliptic curve $E/\mathbb{F}_{p^2}$, the $\mathsf{OneEnd}$ problem asks for a non-scalar endomorphism of $E$. By known reductions, solving this problem also solves the supersingular endomorphism ring and isogeny problems. Wesolowski obtained exponent $1/3$ under an assumption on the factorization of a small degree, whereas the previous unconditional exponent was $2/5$. We give a Las Vegas algorithm, analyzed without a smoothness heuristic, with expected time and memory \[ p^{1/3}\exp\bigl(O(\sqrt{\log p\,\log\log p})\bigr) = p^{1/3+o(1)}. \] The algorithm fixes in advance a family of degrees that are products of small primes. Known counting results provide many isogenies of these degrees from curves to their Frobenius conjugates, and a collision estimate shows that the isogenies occur on sufficiently many distinct curves for a random walk to reach one of them. From such a curve, the algorithm splits a degree into two parts, enumerates two lists of shorter isogenies, and matches their targets to obtain an isogeny to the conjugate, whose composition with Frobenius gives the required endomorphism.

A companion paper found that what a defender must block over time has units: bits per period [Takashita, 2026a]. This paper sets it. Hold every crossing in escrow for one period, audit each held item independently with probability r, and revoke the window if any audit catches something. An adversary crossing k times, each carrying c bits, expects kc(1-r)^k, maximised at k* = 1/ln(1/(1-r)), a bound of L(r) ~ c/(er) per window. The bound is a supremum over the adversary's choice, so the scheme may be public; simulation matches it to 7.7 standard errors. It is a rate, not a total: escrow alone still lets the secret assemble in every run. But if the secret decays at a fraction mu of held bits per period, holdings converge to g/mu at any horizon, so an L-bit secret is unreachable once mu > g/L -- an error threshold in Eigen's sense, sharp where the closed form puts it (100% of runs assemble at 0.9mu*, 0% at 2mu*, over 20,000 windows). Deception that needs the adversary to reason badly fails: a surface whose names lie left accuracy at 18 of 18, and 100% at three reader strengths. Withholding reference works, and differently: no reader would commit at all. Keying the entry points hides 0.10 bits of what a module does; keying the denotation hides 2.64 of 3.00 at chance accuracy, while 100% of readers still call it ordinary Python. Variance must be removed from the audit rate, where loot is convex in r, and added to the activation budget, where survival is multiplicative: extinction 70% to 100% at a fixed mean. End to end the stack takes the leak from 100,000 to 59 bits, a factor of 1,704, leaving 12% of legitimate work standing; keying the window to the caller restores that to 100% at no cost in leakage, at the price of a bound that is per principal. Of 65 read-only tools, escrow leaves 2,400 bits per call: a factor of 10, not infinity.

The importance of deep neural networks (DNNs) is widely recognized, and the parameters obtained through training are regarded as valuable assets. Recently, attacks that extract these parameters using only oracle queries to a DNN have been actively studied at IACR conferences. The hard-label setting is the most challenging setting for model extraction, where an adversary can observe only the final output label, such as "dog" or "cat." At Eurocrypt 2025, Carlini et al. proposed polynomial-time hard-label extraction of ReLU-based MLPs. However, one step of this attack process, i.e., sign recovery, requires a large number of queries and substantial computation. Implementing this step in a black-box setting remains difficult. Consequently, a fully black-box end-to-end demonstration on trained deep ReLU MLPs has remained a challenge. In this paper, we propose a new sign-recovery algorithm based on a completely different principle from the existing method. Our method requires no dedicated queries for sign recovery. In our experiments, it achieves higher sign-recovery accuracy than the existing method. Consequently, it enables efficient sign recovery even for trained models. With our sign-recovery algorithm, all steps of hard-label model extraction can be implemented in a black-box setting. By combining these implementations, we demonstrate end-to-end model extraction from models trained on MNIST and Fashion-MNIST, with width 16 and 4 or 6 hidden layers, achieving over 98% label agreement.

Periodic inspection of deployed wireless devices is necessary because unauthorized hardware replacement may preserve communication functions, credentials, and logical identity, making anomalous devices difficult to detect. Such inspections are conducted under controlled measurement conditions to verify that each device remains consistent with its enrolled hardware state. Conventional radio-frequency fingerprint (RFF) may provide insufficient separation when replacement hardware closely resembles legitimate hardware, while a polarization fingerprint (PF) constructed at one observation direction may miss spatially nonuniform polarization changes. This paper proposes the spatio-frequency polarization fingerprint (SFPF), which jointly represents complex polarization responses over multiple frequencies and observation directions; conventional PF is its fixed-direction slice. We derive SFPF formation from hardware-dependent modal excitation, directional far-field radiation, and polarization projection. A first-order sensitivity analysis shows that the response to the same hardware change varies with both frequency and direction, motivating joint spatio-frequency acquisition. Electromagnetic simulations confirm the nonuniform spatio-frequency sensitivity and show that, under the same observation budget, SFPF improves normalized distance, Fisher score, and the inter-/intra-class ratio over PF by 17.7%, 45.8%, and 11.3%, respectively. Experiments show that SFPF consistently outperforms RFF and PF over 0--20~dB. At 15--20~dB, SFPF achieves anomalous-device F1 scores of 87.3--90.4% and AUROC values of 85.4--95.5%.

Large language models (LLMs) have achieved state-of-the-art performance across a wide range of tasks, motivating two important aspects of deployment: inference efficiency and output provenance, which can be tackled by speculative sampling and watermarking, respectively. However, recent works have shown that combining these two goals is highly nontrivial and can be potentially impossible. In this work, we develop a novel multi-draft speculative sampling algorithm based on Poisson processes that improves the frontier of this fundamental trade-off. The proposed algorithm has strong sampling efficiency on its own and, more interestingly, is naturally watermarkable: we can embed an unbiased watermark without degrading speculative acceptance. Moreover, our algorithm is based on an exact list-coupling-without-communication scheme, which yields a drafter invariance property that benefits both sampling and watermarking. It is the first multi-draft, drafter-invariant speculative sampling scheme that maintains both watermark strength and sampling efficiency, and we experimentally verify its strong performance in both aspects.

Defenses against jailbreak attacks on Large Language Models (LLMs) operate at different pipeline stages, such as input modification or output guard, but it remains unclear which defenses to deploy at each stage and how to combine them. Prior empirical studies, fragmented by inconsistent attack-success-rate definitions and experimental settings, have evaluated defenses largely in isolation. Here we present the first systematic study, to our knowledge, of defense combinations both within and across pipeline stages, under a consistent threat model of direct, black-box, single-turn attacks. Our decision framework standardizes evaluation through a principled attack-success-rate formulation with controlled query budgets, together with explicit fairness rules. Across 19 attacks and 15 defenses, we find that no single defense is universally best, but well-chosen combinations achieve substantial safety with minimal utility degradation, yielding practical recommendations for layered defense pipelines.

We present an architecture that enables data owners to combine private data into data pools using Trusted Execution Environments (TEEs) and manage these pools by issuing narrowly scoped computational rights, encoded as Digital Rights Tokens (DRTs), to third-party data analysts. Each DRT binds specific open-source code to a pool and is issued and redeemed on a distributed ledger. Data analysts can obtain the right to execute open-source code on the combined sealed data inside a TEE and receive the result from this code execution, but not the underlying data. We argue for a control-centric view of privacy in which creators retain ex ante control over how their data is processed. A reference implementation runs WebAssembly (WASM)/Python jobs over sealed datasets and records redemptions on Solana, illustrating the feasibility and limitations of the platform.

Cybersecurity risk quantification remains challenging due to limited operational data and difficulties in quantifying Loss Event Frequency (LEF). This paper introduces the Loss Event Frequency Security Analyser (LEFSA), a probabilistic framework that reformulates LEF estimation as machine-level Cyber Loss Event (CLE) prediction combined with hierarchical infrastructure-level aggregation. LEFSA estimates calibrated machine-level CLE probabilities from operational cybersecurity telemetry and aggregates them across infrastructure layers while accounting for machine-level dependencies. This provides a foundation for scalable, explainable, and operationally applicable cyber risk estimation at the level of machines, services, business processes, and the entire organization. The framework was evaluated using proprietary Managed Detection & Response telemetry from 23 organizations using Microsoft Defender for Endpoint. XGBoost achieved the strongest predictive performance, with a mean area under the receiver operating characteristic curve of 0.90 and consistently low calibration error across evaluation periods. The results demonstrate that operational cybersecurity telemetry contains substantial predictive information for future CLE occurrence, supporting probabilistic machine-level modeling and hierarchical aggregation as a promising foundation for quantitative, data-driven cyber risk management.

Edge AI accelerators are increasingly deployed in safety-critical environments, where model outputs may control physical actuators, make access-control decisions, or trigger alarms. In these settings, runtime failures often remain undetected because model corruption, distribution shift, and adversarial inputs can still produce well-formed, confident predictions. This paper presents TERMon, a lightweight hardware runtime monitor that detects such anomalies by observing inference behavior rather than re-executing or formally verifying the model. TERMon represents class-conditional trusted behavior as hardware-efficient ternary patterns that are matched in parallel against a thermometer-encoded fingerprint. The ternary encoding reproduces the corresponding unquantized range decision exactly. TERMon detects harmful weight corruptions in proportion to their behavioral impact, while out-of-distribution and adversarial inputs are largely not separable using the monitored features at a strict false-positive operating point. We implemented TERMon on a PYNQ-Z2 FPGA, and the pipelined design requires no on-chip block RAM or DSPs and has a two-cycle decision latency.

Privacy leakage in LLM agents is commonly evaluated within individual components such as memory, retrieval, or tool-use pipelines, which makes it difficult to distinguish internal exposure from information that an external observer can actually recover. We present CIPL (Channel Inversion for Privacy Leakage), a channel-aware evaluation framework for black-box privacy leakage in LLM agents. CIPL represents a target through sensitive source, selection, assembly, execution, observation, and extraction stages and evaluates the transition from selected sensitive units to attacker-recoverable output under a shared protocol. Experiments across memory-based, retrieval-mediated, and tool-mediated targets, together with a BrowserUse live-agent case study, show that storage labels alone do not determine recoverability. Memory targets form a near-saturated reference case, retrieval-mediated leakage is frequently partial, and tool-mediated and live-agent leakage varies strongly with observation surface, prompt-to-channel alignment, retrieval depth, and provider behavior. A stratified semantic audit further identifies attacker-useful disclosures that canonical exact matching misses. CIPL therefore provides a common framework for comparing how internal sensitive dependence is realized as externally recoverable leakage across heterogeneous agent pipelines.

Retrieval-Augmented Generation (RAG) improves large language models by grounding outputs in external knowledge sources, but this dependency also creates a surface for poisoning attacks. This paper introduces Micro-Collaborative Poisoning, a distributed attack in which a false target claim is divided across multiple locally plausible documents instead of being concentrated in a single malicious passage. We evaluate the attack across 108 RAG configurations by varying dataset, retriever architecture, retrieval depth, database composition, number of poisoned databases, and generator model. The results indicate that Micro-Collaborative Poisoning is not driven by a single dominant poisoned passage, but by the accumulation of weak adversarial signals across retrieved sources. Increasing top-$k$ and poisoning multiple databases make it more likely that these signals will appear together in the retrieved context, while clean database diversity and stronger retrievers can reduce their influence. The document-level poisoning visibility analysis further shows that this threat is difficult to expose through isolated document inspection, since Micro-Collaborative Poisoning achieves downstream influence while leaving a weaker explicit poisoning signature than direct poisoning.

Recent generations of Apple MacBooks embed an inertial measurement unit (IMU) within their unibody chassis for device orientation and motion sensing. However, this IMU inadvertently captures not only intended device-level information but also subtle physical vibrations from user interactions and the surrounding environment. These signals establish a novel, previously unexplored side channel. We uncover a vulnerability allowing non-root access to IMU data via an IOKit driver, alongside two content-free system metadata interfaces (HIDIdleTime and CGEventSource) that further enrich the side-channel leakage. Through rigorous characterization of the IMU data, we reveal that the leakage spans three core dimensions: (1) keystroke identity (which key is typed), (2) desk surface (where the laptop is placed), and (3) user behavior (who is typing). Leveraging these findings, we introduce BRUTUS, the first comprehensive unprivileged side-channel attack targeting built-in IMU sensors on Apple MacBooks. BRUTUS achieves a character-level accuracy of 89.1% to 97.5% in key recovery. Furthermore, aided by language models, it can successfully reconstruct certain sentences with 100% accuracy. For user identification and environment profiling, BRUTUS correctly discovers user and environment profiles without labels and correctly assigns subsequent segments to their corresponding profiles. Ultimately, this work highlights the urgent necessity of strictly regulating access to built-in IMU sensors.

In cryptography, critical sets of Latin squares have particularly been implemented to design secret sharing schemes. A main problem in these cryptographic protocols arises from absent holders of pieces of information that are common to different critical sets, because they become indispensable to recover the secret. This paper solves this problem by making use of the orbits of entries described by the autoparatopism group of the Latin square under consideration. To this end, we introduce the more general problem of computing critical sets of Latin squares having a given paratopism in their autoparatopism group. These critical sets depend only on the conjugacy class of the autoparatopism and the main class of the Latin square under consideration. Based on this fact, as an illustrative example, we determine the smallest and largest sizes of critical sets associated with autoparatopisms of Latin squares of order up to six. We implement this approach in the design of a new secret sharing scheme.

Third-party adapters for open-weight language models ship as opaque weight matrices; a recipient cannot check whether an adapter hides a backdoor without trusting the publisher or inspecting the weights, the publisher's core asset. For one important class (payloads placed where a safety monitor is structurally blind), detection is unsound as a defense: every detector that factors through the declared monitor is invariant on its blind subspace, and honest and backdoored adapters overlap on every blind-subspace statistic we evaluate, because benign adaptation uses that subspace too. Rather than detect this channel, we make it structurally \emph{absent} and prove that we did. The publisher builds the adapter to read the input only through directions the monitor covers and proves this in zero knowledge, revealing nothing about the read factor it certifies. The certificate is cheap because the expensive part, identifying the monitor's blind spot, is a deterministic function of the \emph{public} base model, so only one linear identity is proved; the served residual is the base model's own public floor, not a prover-chosen tolerance. The result is \emph{ServeGuard}, a supply-chain primitive: the publisher ships a \emph{proof-carrying adapter} whose proof lets a consumer or regulator verify, without the certified read factor and without trusting the publisher, that the adapter carries no hidden channel of this class relative to the declared monitor; an admission-time typing guard binds the guarantee to the adapter bytes admitted at serving time. Across eight checkpoints up to 7B from four families, the monitoring budget is architectural: the measured frontier saturates at the value-path rank on grouped-query checkpoints but not on multi-head ones. On a 0.5B model confinement is nearly free for benign adaptation, making monitor quality the security lever.

Homomorphic encryption (HE) has emerged as a promising approach to privacy-preserving machine learning (PPML), enabling computation directly over encrypted data. In HE-based PPML, a client submits an encrypted input to the server, which evaluates models such as large language models (LLMs) without access to the underlying plaintext. However, we identify a critical security vulnerability in this setting: HE-LLM inference is vulnerable to malicious clients that submit adversarial prompts, such as jailbreak attacks. The same confidentiality that protects benign clients also prevents the server from inspecting incoming prompts or generated responses, making adversarial attempts difficult to detect or block and potentially allowing successful attacks to remain entirely invisible to the server. To address this vulnerability, we propose HE-Guardrail, a framework that evaluates guardrail mechanisms entirely over encrypted data and homomorphically controls whether the target-model response is returned to the client. We instantiate HE-Guardrail with three representative guardrails - Llama Guard, JBShield, and GradSafe. Our results show that HE-Guardrail closely reproduces the decisions of the corresponding plaintext guardrails in the encrypted domain, with distinct security-efficiency-utility trade-offs.

Ransomware increasingly fragments its file operations across temporary and intermediate files, scattering the semantic context that links individual I/O events to an overarching encryption campaign. This fragmentation defeats existing detectors that reason over isolated file streams -- whether pattern-based methods that match rigid event sequences or learning-based methods that require accumulating statistical evidence across many files. We present DEFEAT, a framework that reconstructs this fragmented, scattered context by grouping causally related file events into File Event Gadgets (FEGs), semantically coherent units that capture the full intent behind sequences of file operations spanning multiple dynamically created files. Unlike provenance graphs (system-wide causal graphs that record relationships among all OS entities, such as processes, files, sockets, and registry keys, across the entire system), FEGs are scoped to the file-operation context of a single user asset, enabling lightweight, targeted analysis without whole-system instrumentation. Each FEG is modelled as an attributed control flow graph (ACFG) and embedded via a graph neural network for unsupervised clustering, enabling analysts to label entire behavioural clusters rather than individual samples, reducing annotation effort by 94%. Evaluated on a corpus of 97,816,471 file I/O events spanning 67 ransomware families, DEFEAT achieves 99.2% detection accuracy and outperforms state-of-the-art methods including UNVEIL, RWGuard, and Peeler by 6.57 to 7.56%. The framework operates at the granularity of a single file encryption: because each ACFG represents exactly one FEG (one user asset context), a cluster label can be assigned as soon as the first file operation completes, enabling detection at the first encrypted file.

Software Bills of Materials (SBOMs) have emerged as a key mechanism for software supply chain governance in enterprise architectures. However, their adoption across organizations remains limited due to concerns about exposing sensitive dependency information. To address this limitation, we propose TrustBOM, a scalable architecture for confidentiality-preserving SBOMs integrated into enterprise CI/CD workflows. TrustBOM enables software providers to attest that specific vulnerabilities or restricted licenses are absent from their software without revealing the underlying dependency graph. This is achieved using zero-knowledge non-membership proofs, which are applied selectively based on consumer-defined policy constraints. The architecture ensures that proof generation scales linearly with the number of asserted constraints rather than with the size of the SBOM, enabling efficient operation in large-scale enterprise environments. Empirical evaluation demonstrates linear performance, with an average proof generation time of 0.9 seconds per constraint on commodity hardware, indicating the feasibility of deployment in enterprise platform ecosystems.