Papers

AI agents can select tools, counterparties, and transaction parameters, yet inference should not itself confer authority to execute a financial action. This study develops and evaluates Authority-Inference Separation (AIS), an intent-centered architecture for bounded agentic finance. AIS treats a financial action intent as the control object: a machine-generated proposal can receive temporary executable authority only after an independent deterministic control plane validates registered agent identity, accountable ownership, mandate and risk-appetite lineage, policy version, state, approvals, and exact economic semantics. Blockchain can then enforce the operational representation of granted authority and record portable settlement evidence, while institutional legitimacy, service delivery, accounting classification, and human accountability remain off-chain obligations. Evaluation combines four-domain instantiation, official BIS and MAS cases, a 48-fixture executable prototype, and a public-ledger observability test. Across 36 synthetic authorization attacks, a direct-agent baseline accepted 36 attack effects, a prompt-policy baseline accepted 20, and AIS accepted none; all three accepted 8/8 admissible fixtures. AIS also rejected 4/4 token replays and 8/8 recipient or rail substitutions, withheld completion in 4/4 service-delivery failures, and populated all 13 defined evidence fields. A test of 1,700 recent Base transactions associated with public x402 facilitator addresses shows that public ledgers can evidence settlement and selected authorization parameters but cannot establish institutional mandate, legal accountability, service delivery, or accounting treatment. AIS and blockchain are therefore complementary: AIS decides whether a specific intent may act, while blockchain can make granted authority bounded, executable, and independently observable.

Differential privacy has become the standard for generating privacy-protected data releases. However, differential privacy does not translate intuitively to disclosure risk. In particular, it remains unclear how much an adversary's belief about an individual's inclusion in a dataset can change after observing a protected release. To address this question, we derive upper and lower bounds on the posterior-to-prior ratios of inclusion beliefs under bounded probabilistic and approximate differential privacy. By assuming a worst-case adversary with all-but-one auxiliary information, i.e., knowledge of all except for one of the participants in a dataset, we obtain bounds that apply to any adversary. Because these bounds may fail with non-zero probability, we study the corresponding failure probability for the Gaussian mechanism. We derive a theoretical upper limit on this probability and compare it with Monte Carlo estimates across a wide range of parameter settings. The observed failure rate is several orders of magnitude smaller than its theoretical upper limit, indicating that the latter is highly conservative. These findings suggest that the inferential privacy guarantees provided by differentially private mechanisms may be substantially stronger in practice than what is implied by the theoretical upper limit.

Recently, large language model (LLM) agents, such as Codex, Claude Code, and OpenClaw, have become capable of planning and executing long-horizon tasks through repeated tool calls. This capability also creates new opportunities for prompt injection. Existing attacks either place the malicious objective in one explicit instruction, making it easy to detect, or distribute the intent across multiple execution stages, making successful completion unreliable. In this work, we propose ECLIPSE, a self-evolving and stealthy prompt-injection framework for long-horizon agentic systems. ECLIPSE combines direct user-prompt injection with indirect tool-side injection through two components. On the one hand, Stealthy Attack Trajectory Synthesis uses a sandbox to generate and iteratively verify candidate tool chains, then renders a verified chain as a natural one-shot prompt to serve as the direct instruction. Then, Tool-Chain Steering transfers this plan to the target environment through Static Workflow Encoding (SWE), which embeds state-transition cues in target-tool descriptions, and Dynamic Trajectory Correction (DTC), which supplies corrective signals when execution deviates from the planned chain. To enable systematic evaluation, we further introduce LASE-Bench, a long-horizon agent-safety benchmark with 120 malicious tasks and 198 unique tools; 96.7% of its tasks make at least five tool calls. The experimental results show that ECLIPSE is highly effective: it achieves up to 96.7% attack success without defense and 69.2% under the common safety filter, exceeding the strongest baseline by 27.5% in the defended setting. Evaluations against representative defenses further show that existing safeguards do not reliably defend it, which raises the need for more effective defenses.

Backdoor attacks pose a serious threat to large language models (LLMs), but existing defenses remain fragmented, failing to pro?vide unified defense against both dirty-label and clean-label attacks. To investigate why such fragmentation arises, we present the first systematic feature-level mechanistic analysis of LLM backdoors using sparse autoencoders (SAEs). Starting from a 2 x 2 comparison of clean and poisoned models on clean and triggered inputs, we trace backdoor-induced logit shifts to high-contributing SAE features and categorize them into four roles: interac?tion, suppressed, mixed, and weight-modified features. This taxonomy reveals system?atic encoding differences: dirty-label back?doors are dominated by isolated interaction features, whereas clean-label backdoors rely more on heterogeneous mixtures of mixed and weight-modified features. These differ?ences explain why existing defenses remain fragmented across attack paradigms. We val?idate this hypothesis through inference-time feature clamping, which reduces ASR to at most 10.8% in most dirty-label settings and at most 15.4% in the majority of clean-label settings, while preserving benign-task perfor?mance. These results show that SAE-based analysis can explain defense fragmentation and guide interpretable backdoor mitigation.

Multi-agent applications delegate work across independently operated deployers. After an incident, a verifier must answer two questions: which deployer released the reported bytes, and whether each cross-deployer edge was authorized. Credentials establish who may act, but need not bind them to later output bytes or prove both deployers authorized a dynamically created edge. We present a two-layer attestation design for dynamic delegation without a shared authority, public log, or precommitted workflow. A trusted deployer runtime signs a hash of each released output; this records released bytes but does not prevent prompt injection. Ancestry evidence records edge authorization. Under a unified threat model, we compare a signed linked list, a Merkle-chain variant, and a co-signed DAG. The primitives are standard; the contribution is deployer-side binding and the evidence needed for the two questions. After child-key compromise, the single-signer designs permit an unauthorized parent binding, whereas the co-signed DAG rejects it because the parent must authorize the edge. Fixed adversary matrices and regression tests validate the composed verifier. On an Apple M1 Pro, ancestry-only checks take 24.3-499.2us per hop. In a live local multi-service workflow, a parent discovers the child's A2A Agent Card; the child calls an MCP tool and releases local-LLM output: all 30 signed-DAG tasks passed complete verification, while a controlled child-key-only claim was rejected; its mean end-to-end latency was 813.1ms versus 770.8ms without evidence. In a complementary three-availability-zone AWS deployment, all 1,000 valid co-signed-DAG paths verified; issuance averaged 3.651ms and complete verification 5.015ms. The cloud result excludes TLS/mTLS, KMS, and model-serving latency.

We present how Hyper-V sockets can be used as a real-time communication channel for a malware analysis sandbox. We show that, compared to WinSock TCP sockets, Hyper-V sockets are not subject to TCP/IP-layer blocking and are not enumerated by common TCP connection listing tools. We compare the throughput of the two communication channels as a function of buffer size.

Function secret sharing (FSS) has become a core primitive in privacy-preserving computation. However, each FSS invocation requires a fresh pair of function keys generated by a trusted dealer , expands the system's trust boundary and hinders practical deployment. Existing dealerless protocols eliminate this dependency, but incur substantial communication and a number of interaction rounds that grows linearly with the input bit-width, making key generation a major bottleneck. This paper present KORD, a protocol–hardware co-design that dramatically reduces the cost of dealerless FSS key generation. At its core is a pair of special-purpose chips that establish a common root of trust through mutual attestation and, within it, reconstruct FSS keys—eliminating the need for a dealer. This root of trust further forms a security boundary within which KORD restructures the generation protocol, collapsing the interaction of prior dealerless protocols into a single round, independent of GGM depth. A cross-key scheduling scheme then interleaves independent GGM-tree traversals, sustaining high computational throughput. KORD reduces per-key-generation communication by 7,633–70,274$\times$ over the state-of-the-art distributed FSS protocol across a comprehensive suite of FSS building blocks. Post-route analysis projects 12.75 million 32-bit DPF keys per second at 204 MHz using 21.5K LUTs, with 99.8% AES lane utilization. On private ResNet-18 inference, KORD cuts the share of end-to-end time spent on key generation from over 96% to 11.9%.

Industrial Control Systems (ICS) are the backbone of many critical infrastructure sectors; however, their growing level of connectivity, long lifespan and integration with the Information Technology (IT) environment introduces numerous cybersecurity challenges. The merging of Operational Technology (OT) and IT along with the deployment of Industry 4.0 technologies increases the attack surface of ICS environments, which in turn makes them more vulnerable to advanced cyber threats. Therefore, many researchers have shown interest in the field of cybersecurity of ICS. The topics of intrusion detection, anomaly detection, threat intelligence, attack simulation and resilience assessment of ICS have received much attention. Nevertheless, the development and testing of cybersecurity solutions for ICS remains to be challenging due to the lack of appropriate datasets and experimental environment. The main objective of this paper is to provide the roadmap of existing ICS cybersecurity datasets, testbeds and digital twins. This paper presents various taxonomies along with systematic analysis of architecture, characteristics, capabilities, pros and cons of these tools. The results of the analysis demonstrate the presence of persistent problems such as lack of standardized benchmarking datasets, lack of modern attack scenarios, insufficient number of datasets based on real operational traffic and difficulty in validating artificial intelligence-driven cybersecurity solutions. In addition to summarizing current research on ICS cybersecurity datasets and testbeds, this roadmap provides the identification of research gaps and recommendations on creation of new tools.

Speech enhancement models are widely deployed as frontend modules in real-time speech services, yet their vulnerability to backdoor attacks remains unexplored. Existing backdoor methods are confined to classification tasks and rely on active trigger injection, an assumption incompatible with the passive processing nature of speech enhancement models. In this paper, we propose Ouroboros, a novel backdoor attack framework that leverages the ideal clean outputs of speech enhancement models as natural triggers, enabling inference-time activation without any external trigger injection. Extensive evaluations show Ouroboros achieves near-perfect attack success rates with minimal performance degradation on diverse models and datasets. Physical-world validations confirm that naturally recorded, unaltered clean audio can reliably activate the backdoor. Moreover, Ouroboros generalizes to targeted content-tampering attacks and remains effective against common filtering and finetuning defenses.

LLM agents commonly use knowledge-based tools and access their underlying files, databases, and search indexes through tool invocation. This integration improves agents' ability to provide domain-specific services but also introduces the risk of tool-mediated knowledge extraction: source content exposed to an agent for legitimate responses may be progressively recovered from its outputs, enabling reconstruction of the knowledge source behind a target tool. This paper systematically investigates this risk and identifies two challenges introduced by tool invocation: tool-selection uncertainty, where an agent may invoke a competing tool instead of the target tool, and tool-argument compression, where fine-grained query information may be lost when the agent generates tool arguments. To tackle these challenges, we propose ToolSiphon, a query-only extraction attack that introduces two complementary signals: a target-discriminative signal, implemented through Tool Contrastive Analysis, to steer queries toward the target tool; and a response-grounded factual signal, implemented through Evidence Chained Feedback, to mitigate argument compression and progressively expand extraction coverage. Across three types of knowledge-based tools and six domain-specific datasets, ToolSiphon recovers 74.3% of source records on average when coarse-grained information about non-target tools is available, with 83.2% textual recovery and 90.2% semantic similarity. Even without such information, it recovers 66.3% of source records. ToolSiphon also remains effective against representative defenses and on three real-world agent platforms.

Computer-use agents (CUAs) are agents powered by vision-language models (VLMs) that perceive a screen and operate an operating system through mouse, keyboard, and terminal interactions to automate everyday digital tasks. Their exposure to untrusted content creates a risk of indirect prompt injection (IPI), where an adversary embeds instructions in content the agent reads to redirect it toward actions that violate the user's intent. Evaluations based on fixed, hand-written injections may underestimate the risk posed by adaptive adversaries. We present SIR, a black-box self-improving IPI framework that (i) composes task-specific injections from a small library of reusable red-teaming principles stated in plain language and (ii) uses an iterative feedback loop to analyze unsuccessful attack trajectories and distill new, named principles that are retained in a shared library and reused across tasks. We target operating-system-level compromise and evaluate outcomes through deterministic checks on filesystem, service, and permission state rather than an LLM judge. An attack counts as successful only when both the adversarial objective and the benign user task are completed in the same execution. We evaluate three frontier CUAs, allowing SIR up to 10 attack attempts per case. It achieves joint attack success rates of 24% on Claude Opus 4.8 and 28% on Gemini 3.5 Flash, compared with 4% and 0%, respectively, for the benchmark's fixed, hand-written injection. The red-teaming principles discovered against one victim also improve attacks against other victims, including a different model family, without additional feedback.

Long-term memory is becoming a core capability of LLM agents, enabling personalization and long-horizon interaction. However, memory mechanisms that retain, transform, or expose more information can affect both benign utility and susceptibility to memory poisoning. Existing evaluations typically measure memory utility or attack risk in isolation under fixed configurations, providing limited insight into how stage-specific design choices reshape their trade-off. We present MemGauge, a controllable framework that separately varies writing admission, management policy, and retrieval exposure under matched clean and poisoned conditions. Across 11 LLMs and two long-term memory benchmarks, controlled evaluations reveal three distinct profiles: a threshold-like risk transition during writing, policy-dependent local decoupling during management, and coupled growth of utility and risk during retrieval. We further apply analogous stage-level measurements to four existing memory systems and observe diagnostic associations qualitatively consistent with these profiles. These results show that targeted poisoning risk varies across memory operations and motivate stage-aware evaluation and control of LLM-agent memory.

Serverless platforms based on Confidential Virtual Machines (CVMs) have been recently proposed to address the privacy problems with serverless functions, while achieving low latency. Unfortunately, our study indicates that to achieve these properties, existing proposals impose non-trivial requirements in terms of infrastructure changes and platform memory. Reducio is an alternate serverless platform design that does not require infrastructure changes and significantly reduces platform memory requirements. The platform is designed using two key components: (1) a function isolation framework inside a CVM based on kernel deprivileging features that minimize infrastructure requirements, and (2) a layer-wise caching methodology and algorithm that effectively uses a small in-memory function cache. Our evaluation indicates that Reducio can significantly reduce both platform requirements for deployment and function memory consumption.

Preference optimization is widely used to align large language models with human preferences, but preference-data composition may also influence privacy-relevant memorization. We examine whether adding synthetic privacy-preference pairs to Direct Preference Optimization (DPO) is associated with lower canary-based memorization signals without modifying the objective or introducing a formal privacy mechanism. We propose Privacy-Pressure Preference Mixing (P3M), a data-composition protocol that varies the amount of privacy-preference data while keeping helpfulness and harmlessness preference data fixed. We evaluate a non-privacy Baseline and privacy-mixing ratios of 0.5, 1.0, and 2.0 using Gemma 3 270M-IT across five random seeds and validate the same four conditions using 4-bit-quantized Gemma 2 2B-IT across three seeds. Overall, under the tested conditions, privacy-preference mixing is associated with lower mean canary suffix log-likelihood proxy values across both model settings and lower aggregate membership-inference attack performance relative to the Baseline in the mixed-source 2B evaluation. Specifically, across the privacy-aware 2B configurations, the mean area under the receiver operating characteristic curve (AUROC) ranges from 0.596 to 0.629, and the mean area under the precision-recall curve (AUPRC) ranges from 0.541 to 0.575, compared with 0.804 and 0.790, respectively, for the Baseline. However, the reduction in membership distinguishability does not hold uniformly across data sources. Moreover, the relationship between the privacy ratio and harmlessness preference accuracy varies by model setting, whereas helpfulness preference accuracy remains broadly stable. These findings suggest that P3M should be viewed as a lightweight empirical protocol for examining privacy-utility-safety trade-offs rather than as a formal privacy guarantee or a defense against extraction attacks.

Candidate generation for open-source vulnerabilities is no longer scarce. AI-assisted code review now produces defect candidates cheaply, and industry programs pair them with expert human triage. The remaining scarcity is validation and impact assessment, and the gap is largest in critical-infrastructure software like 5G cores. Here, validation has four costs: verification, reachability, impact, and fix-completeness. We present Drishti, an AI-led human-directed vulnerability audit framework with four components, one per cost: (i) an anti-pattern catalog for verification, (ii) critical-path triage for reachability, (iii) concentric validation for impact, and (iv) patch-review for fix-completeness. Across audits of Open5GS and free5GC, Drishti produced three findings. The first is a pre-authentication NULL-dereference in the Open5GS NRF multipart parser, fixed upstream with a CVE requested. The second is an ASN.1-PER memory amplification in the free5GC NGAP decoder. A 2-byte input from a rogue gNodeB OOM-kills the AMF in 6.2 seconds. The third is a defective patch on CVE-2025-69248 whose defense-in-depth check is dead code before authentication.

Deep learning-based side-channel analysis has historically focused on single-byte targets and manually cropped traces, which risks discarding exploitable leakage. While recent work has proposed specialized architectures and resampling techniques to address this gap, the literature lacks a simple transformer baseline for simultaneous full-key attacks on uncropped traces. We present an open-source transformer implementation for uncropped full-key attacks which uses the standard transformer encoder backbone, adapting only the input and output layers to the side-channel setting. We release our implementation, training recipes, and pretrained weights for uncropped ASCADv1f, ASCADv1r, and CHES-CTF-2018 which achieve performance competitive with previously-reported results, while using less than 10GB of VRAM and requiring at most 3.34 hours of training on a single NVIDIA A6000.

Social virtual reality (VR) creates a paradox. A user's body motion is a high-entropy biometric: head and hand trajectories alone re-identify users among tens of thousands with over $94\%$ accuracy, so anonymizing the rendered avatar is a practical necessity. Yet a user often still wants to prove their identity to a chosen party from inside that anonymity. We present MoSign, which recasts digital watermarking as a challenge-response authentication protocol on the motion channel. MoSign embeds a time-varying keyed message into the style latent of a motion variational autoencoder via keystream-whitened Gaussian-Shading: watermarked motion is provably indistinguishable from watermark-free motion, since any detector's advantage reduces to breaking a pseudorandom function, so the mark composes with anonymization. The message is a keyed MAC over an epoch counter, a session nonce, and a deployment context, making MoSign replay-resistant and bounding forgery by the verifier's measured false-accept rate times the adversary's online query budget. A key-holding verifier decides with a sequential test. We identify render$\rightarrow$record$\rightarrow$re-estimate ("recapture") as the realistic VR attack surface: a generic pose estimator strips the necessarily subtle watermark, but a recapture-robust keyed reader recovers it (up to $0.96$ codeword accuracy on a projected-2D channel, $0.81$ through a full render-to-video loop), while without the key recovery stays at chance. On HumanML3D, MoSign authenticates every legitimate user at a false-accept rate of $10^{-4}$ on clean and most channels and stays undetectable (detection AUC $0.51$, chance $0.5$); on the BOXRR-23 VR dataset it carries the mark through a real anonymizer at $0.99$ codeword accuracy and adds no de-anonymization side channel.

Spectrum governance requires rules to be translated into machine-executable decisions while preserving incumbent protection, regulatory authority, and an auditable record of why a decision was made. We present SPECTRA-GOV, a Tri-Layer Adaptive Governance Architecture (TLAGA) connecting international treaty coordination, national adaptive licensing, and real-time enforcement. The work builds on the original reference implementation preserved at v0.1.0-paper and adds a post-audit evaluation layer rather than replacing it. V2 introduces controlled spectral-contention scenarios, corrected geodesic-distance calculation, per-operator selective authorization, paired baseline counterfactuals, policy perturbation, regulatory-change analysis, and causal provenance. Across 10,000 scenarios in each of seven contention classes, incumbent protection was 100.00% in the no-contention control, 99.92-99.87% in weak-to-dynamic classes, 99.47% under strong overlap, and 90.00% in the adversarial close-proximity class. In a paired S3 baseline experiment, selective authorization achieved 100% incumbent protection and 89.43% access opportunity, while the population-wide dynamic baseline achieved 99.91% protection and 99.53% access. Identical results for the dynamic-SAS and SPECTRA-GOV selective variants mean selective admission alone is not claimed as an exclusive algorithmic novelty. The contribution is the integration of policy representation, computational enforcement, auditability, provenance. Enforcement timing was measured in-process, with mean latency increasing from 0.028 ms for one operator to 1.682 ms for 500 operators; these are software benchmarks, not field measurements. The results establish a reproducible computational governance prototype and identify remaining questions before operational or regulatory claims can be made.

Agent payment protocols are emerging as a key transaction layer for autonomous commerce, enabling AI agents to purchase goods and services and execute payments on users' behalf. Unlike conventional payment flows, they distribute user intent, delegated authority, credential use, settlement, and fulfillment across multiple actors and stages, creating security dependencies that no single message or participant can enforce. Yet these guarantees remain largely implicit across evolving specifications, schemas, and reference implementations, with little systematic formal analysis. We formalize four representative agent payment protocols: x402, MPP, ACP, and AP2 in Tamarin. Using a common abstraction of the agent payment lifecycle, we construct source-grounded models that capture each protocol's roles, state, trust assumptions, and lifecycle transitions. Rather than assuming a complete property taxonomy, we use source-backed verification questions and counterexample traces to expose missing bindings, state constraints, and cross-stage correspondences, consolidating them into 18 shared security principles. Across 86 verification cases, our analysis reproduces 46 known or calibration cases and identifies 40 previously undocumented formal-consistency findings. For each retained violation, we isolate the missing protocol relation, construct a minimally strengthened reference model, and reverify the intended property. We further evaluate the new x402 findings across three implementations and validate ten representative findings through implementation PoCs, SDK/schema-level witnesses, and source-aligned executable traces spanning five security principles. Our results show that delegated authorization must remain consistent with its resulting economic and service effects across actors, states, and protocol stages.

Commercial LLM APIs advertise a specific foundation model, but the served backbone may be silently substituted, quantized, or wrapped, for example to save deployment costs. All existing audits decide backbone identity from the text-output channel, which is structurally fragile for agentic APIs because modern serving stacks (OpenAI, Anthropic, Gemini, Cloudflare Workers AI, LangGraph) discard text and expose only structured actions when the model calls a tool, and provider-injected system prompts can distort text distributions enough that text-channel tests falsely accuse honest providers of substituting the claimed model. We observe that recent agentic post-training internalizes tool-use directly into the weights, opening a new audit channel that the serving stack still exposes and that is largely invariant to deployment context. We introduce Agentic Provenance (AgentProv), the first action-based identity audit for agentic LLM APIs: AgentProv fingerprints a deployed model through its categorical tool-call distribution and decides identity via an MMD permutation test. AgentProv catches every substituted model (100% on 630 evaluated checkpoint pairs), while holding the false-positive rate under system-prompt injection at 7% (vs. 67% for MET and 53% for RUT). On third-party API endpoints, AgentProv's disagreements with MET are consistent with an independent token-count side-channel that detects provider-injected system prompts.

Multi-agent AI platforms move quickly from staging to production, but the way agents establish trust remains rudimentary: an agent either transmits raw data to a peer or accepts that peer's natural-language self-report that a value complies with policy. The first over-shares; the second is unverifiable and is exactly the channel prompt injection attacks. Prevailing responses emphasise identity, visibility, and post-hoc detection, and recent proposals for cryptographically enforced agent policy have been evaluated in simulation rather than execution. We take provable data minimisation between agents from proposal to running system. In our Zero-Knowledge Proof Gateway, agents exchange proofs of governance-defined predicates over private data rather than the data itself, so exposure is prevented by design rather than detected afterwards; because no interoperability protocol can carry such a proof, we propose a slot and implement it on both MCP and Agent2Agent from one endpoint. A 32-bit threshold predicate proves in 6.2 ms and verifies in 1.0 ms with a 608-byte Bulletproofs proof on one commodity vCPU; eleven adversarial experiments and nineteen protocol checks pass; and the system is deployed to Kubernetes with empirically verified network isolation. Our case study proves a retail client order is within its limit without revealing the amount, instantiating the GDPR data-minimisation principle as an enforced technical measure of the kind EU law now names explicitly. We then address the limitation no comparable work resolves: a predicate proof binds a statement to a committed value, never to the system of record. We give a construction fusing an enclave attestation with the proof in both directions, so verifying one artifact certifies jointly that the predicate holds and that the value was read by a specific measured binary, and test it against a mock authority.

Large language model agents place outputs from external skills into their execution context, allowing attacker-controlled data to influence later privileged actions. Existing defenses mainly classify untrusted content or authorize proposed operations. They do not directly address how an agent's future authority should change once untrusted data enters its state. We present SkillGuard, a harness-level enforcement layer that treats this event as contamination and restricts future capabilities to disconnect the resulting state from deployer-defined forbidden states. Given sound skill summaries and policies, SkillGuard represents security-relevant transitions with a Skill Impact Graph, specifies admissible control over skill parameters via steerability signatures, and mediates invocations with an inline reference monitor. Following contamination, it computes weighted capability restrictions using binary, fractional, or fractional-flow strategies without auxiliary language-model inference. We evaluate SkillGuard on four AgentDojo suites with two backend LLMs, Gemini 2.5 Flash and Llama3.3-70B, against an LLM-only No Defense baseline and three defenses at different system layers: Spotlighting, CaMeL, and AttriGuard. We construct a compositional attack benchmark in which each attack combines observations individually insufficient to induce target violation and evaluate the same baselines on it. Under AgentDojo's Tool Knowledge attacks, SkillGuard eliminates attack success on three of four suites for both backends and reduces it to 4.8% and 14.3% on Slack. Against compositional attacks, it outperforms every baseline on Llama and matches the strongest baseline on Gemini at higher benign utility. Fractional-flow restriction preserves substantially more capabilities than binary restriction at the same attack success rate. Across both settings, SkillGuard adds no model calls or token overhead.

System-level mobile agents are evolving from fixed scripts into adaptive systems that continuously observe interfaces, reason, and adjust their actions, allowing automated attacks to navigate dynamic UIs and complete complex tasks. Existing applications detect automation using touch trajectories, action timing, and the physical coupling between touch and inertial measurement unit (IMU) signals. However, a privileged system-level agent executor can control both touchscreen input and application-visible sensor delivery, enabling it to jointly generate time-aligned touch and six-axis IMU signals and evade these defenses. We present ActReal, a physical-action attack framework for system-level mobile agents. ActReal converts semantic agent actions into task-valid touch and IMU events using genuine-trajectory adaptation and physics-guided IMU generation. ActReal achieves a mean event-level attack success rate of 77.5%; even when detectors jointly observe touch and IMU, its attack success rate remains 71.1%.

Post-quantum migration increases WebPKI authentication cost, but authenticating a compressed certificate object does not by itself preserve the mutable authorization context under which a relying party accepts it. We formalize context closure: the authenticated projection accepted by a verifier must determine the selected authorization semantics it claims, relative to declared source contracts and event-coverage witnesses. We instantiate this idea with \LRp, a two-plane post-quantum construction that authenticates mutable CA-context state in an update plane while the warm path carries only state-local dependency references selected by explicit profile negotiation. In a pinned CCADB reconstruction, we obtain 44,912 path/view contexts and 16,858 physical CA lineages across Apple, Chrome, Microsoft, and Mozilla views. The core compiler yields $m_{50}=6$, $m_{95}=16$, and $m_{\max}=18$ typed dependencies. A warm LR+ selector therefore costs 296, 776, and 872 bytes at median, p95, and maximum, compared with 3,842, 5,932, and 6,350 bytes for a one-signature stateless bundle carrying the same dependency vector. The retained all-view closure state is 16.15 MB, and per-view lifecycle crossovers range from 19.60 to 50.41 median-path warm authentications/day under the stated checkpoint and update model. The implementation and evaluation artifact are available at https://github.com/nserser/LR-WebPKI

Enterprise networks remain vulnerable to Advanced Persistent Threats (APTs), where adversaries gain an initial foothold and move laterally across the network, accumulating access permissions hop by hop to reach critical targets. Existing network defenses cannot track user movement at the process level across the network; instead, they grant ambient trust to all processes within a host. As a result, once a host is compromised, malicious processes inherit the victim's permissions, thereby expanding the attacker's access scope and enabling further lateral movement. To address this gap, we present NetZone, an in-network access control that confines each user process to a fixed access scope that persists as the user moves across the network. NetZone introduces a new abstraction, called AccessScope, which represents a lightweight access capability bound to the user's processes. Each AccessScope encodes the set of hosts a user identity is authorized to access and is embedded in the process's outgoing network traffic for validation before reaching its destination. As users pivot across hosts, AccessScope propagates with their traffic, rebinds to the receiving process, and persists across hosts. This ensures that regardless of network location, the user's processes are consistently governed by their bound AccessScope and their access permissions remain unchanged. To handle the high volume of network traffic generated by processes, we develop a data-plane co-design that integrates programmable switches with eBPF. NetZone employs a set of in-network optimizations and lightweight AccessScope persistence techniques to inspect the embedded AccessScope on the fly, enabling line-rate processing of high traffic volumes with negligible latency overhead. Our extensive evaluations show that NetZone can effectively defend against sophisticated attack scenarios without introducing noticeable overhead.

The Learning With Errors (LWE) problem, introduced by Regev in 2005, is central to modern cryptography and post-quantum security. The algorithms to solve the search version of the problem, Search-LWE, can be broadly categorised into algebraic, combinatorial and lattice-based. In this work we propose a new algebraic algorithm for the Search-LWE problem. At a high level, the algorithm combines linear-algebraic techniques with S-polynomial-based methods from Groebner basis computation. We provide a direct complexity analysis of our algorithm, avoiding semi-regularity assumptions and complexity bounds derived from the degree of regularity. Our algorithm achieves a polynomial improvement in complexity over prior results that use Groebner basis methods to solve Search-LWE.

The key limitation of current state-of-the-art influence-based guardrails is that they do not reliably distinguish a legitimate, user-authorized action from a malicious, unauthorized action when both rely on external tool information. This ambiguity can cause benign actions to trigger unnecessary verification and intervention, reducing utility and adding latency. We expose this limitation through an authorization-equivalence audit of 96 conditions derived from 24 base cases. Within matched source comparisons, we hold authorization, the exact committed action, and its intended effect fixed, changing only whether a required value comes from the user or a legitimate tool result. Although the action remains unchanged, this harmless relocation shifts the causal signal toward the attack region in all 24 cases under both Llama and Gemma scorers. Matched unauthorized controls show that the signal remains attack-sensitive, yet the benign relocation produces a larger average score shift than the actual change in authorization. Architecture-level evaluation shows how this mismatch propagates through guardrail designs. With a semantic monitor, attack success is 0% and utility is 28%, compared with 16% and 60% without it. A shadow-based guardrail allows every tested harmless run, yet does not reject matched unauthorized actions more often overall: 57.5% of unauthorized runs pass automatically before reaching the later security check, compared with 29.2% of authorized runs. These results show that the studied causal signal reveals what shaped an action without reliably encoding whether the action was authorized, and that reference construction and routing are integral to the effective security decision.

The impact of Python applications has been evidenced by their widespread presence in some of the most impactful software domains, such as machine learning frameworks and scientific computing platforms. These applications often integrate native code components written in a lower-level programming language like C. This multilingual construction brings various benefits such as greater performance efficiency and easier interoperability with diverse runtime environments. However, bugs in the native code (i.e., native code bugs), which are usually stealthy, also constitute a major additional challenge to the quality of the Python applications as a whole. Yet despite existing relevant studies, there remains a lack of comprehensive understanding of native code bugs in Python applications. In this paper, we aim to mitigate this knowledge gap through the first in-depth study of such bugs, dissecting their common symptoms, introducing locations, manifestation characteristics, root causes, and fixes. Based on our extensive automated and manual analyses of 216 native code bugs in real-world Python projects on GitHub, we obtained novel findings about and new insights into the occurrence mechanisms and resolution strategies of those bugs.

Modern software systems are commonly constructed in multiple, interacting programming languages. This construction leads to additional, often stealthy vulnerabilities buried in complex information flow due to language interactions. Existing static analyzers are impeded by the heterogeneous semantics of different languages, whereas dynamic approaches suffer from the limited coverage of (available and/or generated) test inputs. In this paper, we develop PolyFlow, a neural-symbolic framework for statically reasoning about information flow across language boundaries, combining large language models (LLMs) and static analysis synergistically. Governed by the control-flow representation of a given multi-language system, PolyFlow leverages LLMs to identify implicit flow facts due to challenging language features, hence augmenting the base representation and then propagating data flow through the system. It tackles inherent barriers (e.g., token limit and hallucination) of LLMs by putting them under careful guidance (e.g., static-analysis-guided scoping, context management, and fact checking), along with a multi-LLM expert panel for negotiated validation. Our experiments on real-world Python-C and Java-C systems show that PolyFlow is cost-effective and superior to various kinds of state-of-the-art baselines, revealing previously unknown cross-language vulnerabilities that are missed by all the baselines.

Hardware Security Module (HSM) logs capture security-critical behavior, but anomalies emerge from relationships across event sequences, keys, object states, sessions, and temporal patterns rather than isolated events. Existing methods separate detection from HSM-specific evidence validation and reporting. In this paper, we present HSMLog, a two-stage framework for HSM log anomaly detection with retrieval-grounded behavioral analysis. In Stage 1, a small language model (SLM) identifies candidate alerts from sliding windows of structured HSM events and performs policy-guided assessment using HSM-specific operational rules. In Stage 2, retrieved policies and historical suspicious-key records strictly predating the alert window, together with candidate-related log context, support conservative candidate review and incident analysis. Evaluated on real industrial HSM background logs augmented with anomaly scenarios co-defined with industrial partners, HSMLog achieves 98.97% precision, 96.00% recall, 98.66% anomalous-event coverage, and a 97.46% F1 score, demonstrating effective anomaly alerting and incident triage in the studied setting.