Papers

Matter aims to unify smart home ecosystems through an open, interoperable, and secure standard, relying on cryptographic mechanisms for device authentication and data confidentiality. However, its openness also exposes protocol details, credential structures, and implementation characteristics to adversaries. We show that an attacker with temporary physical access can exploit design and implementation flaws to extract credentials and impersonate devices and controllers. These replicas integrate seamlessly into the fabric, enabling persistent surveillance and control even after the attacker departs. Notably, the attack is vendor-agnostic and requires no device-specific reverse engineering, as long as the device is not physically secured. We validate its practicality through a proof-of-concept on a simulated smart home with both commercial devices and development boards. Our findings uncover previously underestimated attack surfaces in Matter, particularly under physical access scenarios, and motivate four targeted mitigation strategies.

Split learning lets a client train a language model on a server without sending its text. The client runs the first layers itself and sends the server only their output, a vector of numbers for each token. During training, the server sends gradients back. We show that an observer at the split can rebuild most of the client's text from this traffic, and we measure how much the gradients help. On GPT-2, an attacker who holds only the publicly released weights of the client's layers recovers 94.20% of tokens from the activations alone and 97.38% when it also sees the gradients, 3.17 percentage points more 95% interval [2.72, 3.64]. Counted by document, the difference is much larger. The attacker rebuilds 13.71% of 32-token documents exactly without the gradients and 37.77% with them, because a document only counts when every token is right. How we count also changes how good a defence looks. Secret mixup, which blends each outgoing vector with a decoy, stops the attacker from rebuilding almost any document exactly, yet the attacker still recovers 83-91% of tokens. In a second experiment on GPT-2 and Qwen3-0.6B, where the server trains only a run of consecutive layers, the layer at which the run starts changes both model quality and leakage, even when the run's length is fixed. We recommend reporting leakage both per token and per document, and treating what a split model sends as being as sensitive as the text itself.

Synthetic gene expression data is increasingly proposed as a privacy-preserving substitute for controlled-access genomic repositories, but its safety depends on empirical auditing. Membership inference attacks (MIAs) provide that audit by testing whether a patient's gene expression profile was used to train a generative model. We report a red-team study on synthetic gene expression data derived from bulk RNA-seq profiles in The Cancer Genome Atlas, released by the ELSA Health 2026 Challenge. We unify five no-box attacks under a single weighted-distance framework in which each variant differs only in how it weights genes: uniformly (baseline), by variance, by synthetic-versus-reference KL divergence, by spectral residualization removing dominant principal components, and by curated pathway membership. Against a conditional variational autoencoder, spectral residualization raises AUC from 0.8251 to 0.8951 and TPR at 1% FPR from 0.3842 to 0.5970 on pan-cancer TCGA. Biological pathway priors did not transfer across targets.

Dense embedding models are core to modern text retrieval, enabling systems ranging from web search to retrieval-augmented generation (RAG). Yet, retrievers are usually deployed within opaque systems, exposing only ranked results, cited sources, or generated answers. This opacity prevents users from verifying which retrievers providers serve and may create a false sense of robustness against retrieval attacks. We show that attackers can infer retrievers' identity only through queries. We introduce TellTail, a fingerprinting attack for identifying retrievers behind black-box systems across a spectrum of access levels. Despite sharing many properties, retrievers can be steered to emit distinct results. When retrieved passages are exposed, TellTail compares retrieval overlap to deduce the underlying model. When only the final generated response is exposed, TellTail optimizes model-specific queries that induce a chosen retrieval behavior on the target retriever but transfer poorly to others. Interestingly, the poor transferability that limits attacks elsewhere is exactly what makes them useful for fingerprinting. We evaluate TellTail using 53 retrievers under three increasingly restrictive settings. TellTail perfectly identifies the deployed retriever from full retrieval rankings; in 94.3% of attempts with unordered top-3 results; and in 92.5% of cases from language-model-generated answers alone. Furthermore, fingerprinting succeeds against a popular RAG system (OpenWebUI). Overall, TellTail shows that retrieval-based systems leak the identity of their embedding model—exposing intellectual property and enabling reconnaissance ahead of model-specific attacks such as corpus poisoning.

Large language models (LLMs) offer a flexible approach to network intrusion detection, but direct classification of numerical flow records can be unreliable without traffic- specific decision boundaries. Retrieval-augmented generation (RAG) enables LLMs to leverage external experience; however, free-form textual experience requires the model to infer the numerical conditions that distinguish traffic classes. We propose SERA-IDS, a Structured Experience Retrieval-Augmented Intru- sion Detection framework that learns decision knowledge from classification errors. A Small Language Model (SLM) analyzes misclassified flows and generates structured rules containing behavioral descriptions, feature conditions, class-confusion infor- mation, tool-derived confidence, and provenance. A confidence gate admits supported rules into the Experience Library. During testing, the library is frozen, and retrieved class-diverse rules are condition-matched to the query flow and provided as evidence to the decision SLM. The same library is evaluated with three small, locally deployable SLMs, namely Llama 3.1, Phi-4:14B, and Qwen2.5:7B, without fine-tuning or paid hosted inference. On the test sets, SERA-IDS improves macro F1 on NF-BoT- IoT from 11.66%, 9.26%, and 7.83% to 86.46%, 69.50%, and 83.51%, respectively, and on NF-ToN-IoT from 3.82%, 6.17%, and 1.94% to 84.16%, 87.66%, and 83.37%. These results show that structured experience retrieval enables small, free, and locally deployable SLMs to achieve strong intrusion-detection performance, including performance exceeding that of the re- ported existing work used in our comparison.

Space infrastructures play critical roles in modern society, including satellite communications (SATCOM). Like the Internet, space infrastructures are vulnerable to cyber attacks, or space cyber attacks, highlighting the importance of managing cyber risks to space infrastructures, or space cyber risks. However, adequately managing space cyber risks is an open problem. In this paper, we propose the first Space Cyber Risk Management (SCRM) framework to systematically analyze and mitigate space cyber risks. The framework models space infrastructures, space missions, and space cyber attacks, while offering algorithms for space mission risk analysis and hardening. We demonstrate its usefulness by conducting a case study on real-world space cyber attacks against the SATCOM infrastructure implemented in our testbed. Our results show, among other things, that: (i) dealing with space cyber attack cascading effects is essential to space cyber risk management; (ii) the framework can effectively harden space missions; (iii) National Institute of Standards and Technology (NIST) security controls can effectively mitigate space cyber risks.

Instruction fine-tuning adapts a pretrained language model to follow instructions by training it on instruction–response pairs from a collection of tasks, such as summarization and question answering. Task-level poisoning exploits this task structure to manipulate the fine-tuned model into producing attacker-specified biased content on a particular target task, without requiring an explicit input trigger. Detecting such attacks is challenging because there is no explicit trigger to identify, the target task and biased content are unknown, and benign fine-tuning itself changes model behavior. We introduce Localize-and-Detect, a two-stage black-box auditing method for task-level poisoning that requires only outputs from both the base and fine-tuned models. In the first stage, we localize the target task by identifying candidate tasks on which the fine-tuned and base models have the largest differences in their next-token distributions. In the second stage, we search the shortlisted tasks for biased content that repeatedly appears in the fine-tuned model's responses but not in those of the base model. We evaluate Localize-and-Detect on 216 poisoned models across two model families, varying the target task, poisoning mode, poison budget, and type of biased content. Our evaluation demonstrates that Localize-and-Detect can effectively localize target tasks and detect biased content across a range of poisoning settings and models, with detection that tracks attack success and few false positives on clean models.

Large language model (LLM) watermarking is important for ownership verification and intellectual property protection. However, existing approaches focus on algorithmic design while treating LLM inference engines as separate components. This separation often introduces auxiliary models or external tools, increasing latency and memory overhead while limiting the use of modern inference optimizations. As a result, a deployment gap remains: practical watermarking must preserve utility, detectability, and robustness while minimizing serving overhead. To bridge this gap, we propose SWIFT, a framework that co-designs LLM watermarking with modern inference infrastructure. SWIFT (i) integrates text generation and watermark construction within a shared LLM backend to reduce re-computation, improve cache reuse, and simplify system complexity; (ii) uses instruction-guided candidate generation with semantic protection to preserve factual consistency and meaning; (iii) executes watermarking concurrently with generation through asynchronous co-serving and adaptive scheduling to reduce latency and improve resource utilization; and (iv) leverages vLLM optimizations to enhance serving efficiency. Extensive experiments across domains and tasks show that SWIFT achieves strong utility, detectability, robustness, and downstream performance while substantially improving serving efficiency. Specifically, it achieves the highest utility score (4.87), 99.65% detection accuracy compared with 76.65% for the low-latency baseline SynthID, and up to 97.7% detection under watermark removal attacks, with 5.9x lower latency than the highest-utility baseline, SafeSeal. These results demonstrate the benefits of co-designing watermarking with inference infrastructure for practical LLM serving.

Advances in large language models (LLMs) have enabled AI-driven code generation from natural language specifications, introducing new attack surfaces for injecting vulnerabilities into software. Prior work has studied this problem only in benign settings where vulnerabilities are introduced inadvertently, or under unconventional threat models where the LLM itself is malicious (backdooring) or the user is the attacker (jailbreaking). In this paper, we study a more realistic threat model: a third-party adversary, with capabilities comparable to existing cybercriminals, compromises the AI code generation pipeline to deliberately introduce vulnerabilities. We call this the EviLLM attack. We have implemented two instances of EviLLM, each of which only requires the underlying LLM to be accessed through a compromised account or browser, and can inject vulnerabilities from 13 CWE classes. As we show in our feasibility study, both attack vectors are already used to implement many existing cyberattacks. Our user study shows that 7 out of 8 and 10 out of 13 participants did not notice the vulnerabilities injected by the two instances of EviLLM, and 13 out of 21 participants "rarely" or "never" considered the risk of an attack like EviLLM.

Design-level security weaknesses can arise from requirements, trust assumptions, missing controls, and data flows before implementation begins. Existing security practices often identify these issues after code is written. We present a multi-agent LLM framework for design-time security analysis from product requirement documents and architecture diagrams. The proposed framework parses architecture diagrams into graph representations, generates misuse and failure cases, constructs attack trees, checks governance and compliance gaps, recommends mitigations, assigns enterprise security-domain tags, and produces a candidate revised architecture recommendation for expert review. The framework does not retrieve from Common Weakness Enumeration (CWE) databases at inference time. Instead, it analyzes system behavior, trust boundaries, component interactions, and data-flow assumptions. Misuse cases act as intermediate representations that link findings to system components and attack paths, while a validation and refinement loop filters unsupported findings and improves grounding, traceability, and actionability. We evaluate the framework on a Microsoft reference-labeled threat-modeling example, labeled synthetic PRD–architecture pairs, and two open-ended systems: Berty and Gas Town. The reference-labeled case supports threat-recovery and actionability analysis, while the open-ended cases evaluate validity, noise, traceability, actionability, redundancy, and attack-tree quality. Results show that architecture-informed, misuse-driven reasoning improves review quality compared with single-shot and ablation baselines. Keywords: LLM Multi-Agent Systems, Design-Time Security, Threat Modeling, Vulnerability Discovery, Architecture Diagrams, Security Analysis, Misuse Case Derivation, Attack Trees, Iterative Reasoning, Security Governance.

AI agents that operate a computer through its graphical user interface (GUI) are being widely deployed. AI control studies how to prevent an AI system from causing harm even if it is misaligned and actively trying to do so. Most control research to date has focused on coding agents, leaving computer use largely unexplored. We introduce Control OSWorld, a control evaluation that pairs 318 tasks from OSWorld with 81 harmful side tasks (e.g., exfiltrating a private file) that an agent must complete without being caught. We use Control OSWorld to create and study control monitors that detect malicious agents interacting with a GUI. We find that a weaker monitor can reliably distinguish honest from malicious trajectories produced by a more capable agent when it sees the full trajectory (97% recall at 3% false positive rate). However, when the monitor has to score each step before it is executed, recall drops at low false positive rates because each action is judged with less context. Monitor performance also depends on what the monitor can see: access to the agent's visible text is critical, while screenshots provide little additional value.

We show that if quantum commitments exist, then either there is no polynomial-time solution to the unitary synthesis problem, or $\mathsf{BPP} \neq \mathsf{NEXP}$. Thus, showing unconditionally that quantum commitments exist would require answering at least one of two longstanding open questions in complexity theory. We prove our main result as a consequence of a more general lemma, which shows that every unitary in $\mathsf{unitaryPSPACE}$ either cannot be synthesized efficiently relative to any classical oracle, or can be synthesized efficiently with an oracle for $\mathsf{NEXP}$ search problems. Our lemma has other noteworthy consequences, including that certain oracle separations involving $\mathsf{unitaryPSPACE}$ would imply breakthrough classical lower bounds such as $\mathsf{NC} \neq \mathsf{NP}$. Along the way, we propose new definitions for the unitary complexity classes $\mathsf{unitaryP}$ and $\mathsf{unitaryPSPACE}$. Our changes address the biggest conceptual issues with definitions suggested in prior work, and lead to elegant proofs. We study both implementations that erase garbage and implementations that allow it, because we cannot rule out the possibility that the two definitions differ. Nevertheless, we show that both definitions can be viewed as special cases of each other. We also showcase many other ways in which our definitions are robust. For example, we show that $\mathsf{unitaryPSPACE}$ has an equivalent characterization as the set of unitary transformations whose entries can be computed to arbitrary precision in polynomial space. Consequently, we deduce that $\mathsf{unitaryPSPACE}$ can generically erase garbage, a result that provably fails relative to unitary oracles.

Federated Learning (FL) is a privacy-oriented learning paradigm that enables collaborative model training while keeping training data local to participating clients. However, it does not guarantee that clients submit policy-compliant contributions or that aggregators process admitted contributions correctly. Existing verifiable FL systems tailor validation rules to specific FL settings, learning workflows, and cryptographic constructions, limiting their applicability across network topologies, participant roles, and aggregation semantics. In this paper, we present PoCoFL, a policy-compliant federated learning framework that separates three aspects: (i) FL type, (ii) policy semantics, and (iii) cryptographic realisation. We provide a formalisation that captures client and aggregation requirements as policy-dependent relations. Clients prove compliance of their contributions using commitments and non-interactive zero-knowledge proofs, while aggregators prove that the recorded set of admitted contributions was processed according to the selected aggregation policy. We demonstrate PoCoFL through four formal instantiations: (i) vanilla, (ii) continual, (iii) personalised, and (iv) threshold-encrypted federated learning. We evaluate the effects of policy enforcement on the learning objectives of vanilla, personalised, and continual FL. We further implement proof-of-concept realisations of all four instantiations, demonstrating the versatility and practical feasibility of PoCoFL. Overall, these results show that PoCoFL can capture complex policy representations while remaining network-topology agnostic.

Quantum fire is a recently introduced cryptographic primitive consisting of efficiently preparable quantum states, called flames, that admit efficient cloning but resist efficient telegraphing, namely reconstruction via classical communication without preshared entanglement. In all prior constructions of quantum fire, cloning is a public operation that requires no separate key and every holder of a flame state can clone it. For applications to access control, however, an issuer may wish to delegate cloning to designated quantum servers while withholding this capability from other flame holders. To address this, we introduce delegatable quantum fire, in which cloning requires a separate key. We give two constructions in the classical-oracle model. Our first construction uses a classical secret key which enables cloning, and any user with the entire key may clone successfully. Our second construction, which we call torch-fire, uses quantum cloning keys, called torches, which enable cloning while remaining unaffected in the process but cannot otherwise be split or delegated to enable additional cloning. An efficient adversary given $m$ torches cannot, except with negligible probability, enable more than $m$ noncommunicating parties to each clone a fresh, independently issued challenge flame. The adversary may jointly process its resources before separating the parties and distribute arbitrarily entangled registers among them. Both constructions are in the oracle model, relying on public classical oracles that allow queries in quantum superposition.

We present a unified framework for upgrading a broad class of cryptographic primitives to support constant-rate certified deletion. Previous constructions require a linear number of qubits per encrypted bit of certified-deletable plaintext. In contrast, we obtain the first constant-rate constructions in the plain model that achieve certified deletion while preserving everlasting security. Our approach applies to a wide range of "all-or-nothing"-type primitives based on BB84-style encodings, including commitment schemes, public-key encryption, attribute-based encryption, and fully homomorphic encryption. Beyond this class, we also obtain constant-rate certified deletion for primitives built from subspace coset states, such as blind delegation, secure software leasing, functional encryption, and differing-inputs iO, and CCA-PKE. Importantly, our framework does not introduce any additional assumptions beyond those required by the underlying certified deletion primitives. Finally, under the hardness of SIS, we show that public verifiability can be incorporated into BB84- and coset-based certified deletion. In combination with our constant-rate constructions, this yields publicly verifiable certified deletion schemes with constant rate.

Security benchmarks for LLM-based agents often report the attack success rate (ASR) as a measure of model robustness and use these scores to compare different models and defense mechanisms, assuming that they describe the security of the agent. In this paper, we explore whether it also influences the benchmark's measurement. To measure the effect of the benchmark representation, we introduce threat-preserving representation sensitivity (TPRS), which measures how much the ASR changes when we change the agent-visible representation while holding the underlying task, harmful action, security policy, ground truth, environment, and the evaluation criteria fixed. On Agent Security Bench (ASB), replacing threat-related tool names with threat-neutral names raises the committed attack success rate by 11.67 percentage points on GPT-5-mini and by 13.21 points on Claude Haiku 4.5. On MCPTox, replacing the original neutral tool name with an explicit threat-related name lowers the ASR by 11.00 percentage points on GPT-5-mini and 4.11 points on Claude Haiku 4.5. On AgentDojo, adding threat-related wording to the attack-relevant tool changes ASR by only 0.50 percentage points on GPT-4o-mini, yet the benign utility falls by 5.36 points on tasks requiring that tool. We ran an experiment on MCPTox where we observed that a threat-neutral name matched on token count, length, and casing reproduces most of the shift produced by the threat-explicit name (8.54 of 11.00 points on GPT-5-mini). The results show that a security score measured under one representation may fail to generalize across threat-preserving representations of the same security problem. Robustness claims should therefore be supported by performance across a controlled set of threat-preserving representations rather than relying on a single representation-dependent score.

Recent studies have shown that most commercial direct time-of-flight (dToF) LiDARs can be spoofed by injecting high-frequency laser pulses into the receiver, which can erase pedestrians from the point cloud. This paper presents the first dToF LiDAR system-on-chip (SoC) with integrated sensor-level hardware security against spoofing attacks. We propose Dual-Domain Fingerprinting (DDF), which emits laser pulse pairs whose time interval and amplitude ratio are both randomized and authenticates received echoes in this two-dimensional space, so that spoofed signals are rejected before they corrupt the ranging result. An Event-Driven AFE (ED-AFE) activates the ADC only around pulse peaks: it digitizes three samples around each peak with a triggered ADC at 1-GHz sampling and applies parabolic interpolation, achieving 1-cm distance resolution with a 99% reduction in ADC power. A time-modulated laser driver controls the laser amplitude from 10% to 100% by modulating the charge time of a laser capacitor, providing the microsecond-order amplitude modulation required by DDF. A LiDAR system with a 16-channel 65-nm CMOS prototype SoC demonstrates up to 120-m ranging and an AFE power of 3.1 mW per channel, 60% lower than prior art. In a proof-of-concept experiment in which the dual-domain authentication is applied to measured sensor data, 73% of the point cloud is protected under spoofing attack, compared with 0% without DDF.

A reasoning model asked whether a causal effect is recoverable from observational data can fail in two ways: it refuses an identifiable query or answers a nonidentifiable one. The latter is more consequential, as no observational data can validate the claimed formula. Measuring this failure requires queries that are provably non-identifiable, which prior evaluations lack, and grading that accepts correct formulas in any equivalent form, which string matching cannot provide. We build CERTID, a formal identification pipeline that addresses both limitations. CERTID uses the sound and complete causal identification algorithm ID to certify whether an effect is identifiable from a given graph and query, and verifies returned formulas against structural causal models whose interventional distributions are known exactly. CERTID further develops theoretical results to mitigate structural leakage, repair non-identifiable queries, and establish grading guarantees. We evaluate three frontier reasoning models (Gemini Flash, Gemini Pro, and GPT5.5) on 1,200 certified instances spanning 4 to 50 vertices. Accuracy proves a poor proxy for soundness: on identical instances, the false-claim rate on non-identifiable queries varies by seventeen-fold across models. We also find that models decide identifiability with 97-100% accuracy on graphs generated after the strongest model's training snapshot. Instances, the certification procedure, the verifier, and per-instance records are available at https://anonymous.4open.science/r/certid-D718.

Static-analysis scanners can identify personal-data types in source code, but they lack mechanisms to connect these findings to standardized privacy vocabularies. PrivDev maps 122 Bearer CLI data types to Data Privacy Vocabulary Personal Data (DPV-PD) categories and links them to potentially relevant GDPR provisions. Our approach combines deterministic mapping for 43 exact-label matches with a retrieval-grounded Large Language Model (LLM) to resolve the remaining 79 non-trivial mappings. The resulting RDF knowledge graph contains 118 ODRL policy resources that were structurally validated using SHACL. The artifact passed five complementary validation gates that cover structural correctness, query consistency, retrieval quality, LLM-based assessment, and human evaluation. In human evaluation, nine annotators produced 711 judgments, yielding a raw agreement of 0.72, Gwet's AC1 of 0.68, and Gwet's AC2 of 0.88. Our results indicate that the proposed mappings are plausible and reproducible, while also revealing ambiguities in scanner-defined data-type labels and coverage gaps in DPV-PD.

Mechanistic edits (ablations, weight edits, activation steering) are the standard tools for unlearning a harmful capability from a neural network while preserving useful ones. Current approaches validate their effects only by testing, which can never cover an entire continuous region of inputs. Prior work at the interpretability-verification boundary certifies descriptions of a model: what a circuit computes, or whether it faithfully explains the whole. We instead certify the behavioral effect of an edit: that disabling a circuit removes one skill and provably preserves another, for every input in a region; a feature non-interference guarantee in the information-flow-security sense. We demonstrate such certified edits from toy ReLU networks up to a standard softmax + LayerNorm transformer, proving removal and preservation over continuous embedding-space regions and reaching roughly 9x the input-perturbation dimension an exact solver can handle by switching to sound bound propagation. Furthermore, we prove that no finite deterministic black-box test can certify removal, exhibiting an edit that passes exhaustive testing yet provably fails on a survivor pocket that can be made arbitrarily small. Guarantees hold on small, standard-architecture networks and, like any removal claim, presuppose that the target skill admits a decidable specification, a property which real-world harms may not have.

AI services increasingly rely on black-box security guardrails, yet privacy-preserving model auditing regimes often cannot measure how well these systems perform in both a human eyes-off production setting, which disallows human inspection of user input, and a machine eyes-off setting, which disallows model inspection of such input. We introduce CorrectGuard, an eyes-off correctness estimation framework for both settings, which involves an independent model-based evaluator predicting whether guardrail decisions on human- and machine-inaccessible inputs are correct using only labeled eyes-on data and without access to the guardrail's internals. We evaluate in-context learning, embedding, and finetuning-based correctness models under leave-one-dataset-out evaluation across 13 safety and security datasets spanning harmful content, jailbreaks, prompt injection, and extraction, and across open-weight guardrails treated uniformly as black boxes. Across both human and machine eyes-off settings (the latter implemented using privacy-preserving fingerprinting of inputs), in-context-learning-based correctness classifiers substantially improve error identification across guardrails, achieving up to a 25 percentage-point increase in macro accuracy, as do finetuning-based approaches which provide a nearly 15-point boost, although performance varies sharply across guardrails and held-out datasets. Correctness scores also support guardrail decision ranking and abstention: across 3 guardrails, the best correctness rankings reduce AURC from unranked baselines of 0.33-0.44 to 0.17-0.22, while the best operating points retain 37.5-52.0% of guardrail decisions at 15% observed risk. These results show that external correctness models can expose systematic failures and support guardrail decision abstention without privileged access to the guardrail.

LLM agents increasingly screen tool outputs with small prompt-injection detectors, and teams choose among detectors by their scores on public benchmarks. We ask whether those scores predict how a detector behaves inside an agent. We replay the ground-truth tool calls of two agent benchmarks, AgentDojo and tau-bench, without an LLM to obtain tool outputs that are benign by construction, label injected outputs by differential replay, and evaluate fifteen detectors, including Meta's Prompt Guard 2, and two task-aware LLM judges on these outputs and on the BIPIA benchmark. Detection rankings transfer poorly between benchmarks: the best detector on BIPIA catches 2% of AgentDojo injections at a 1% false-positive rate, and a detector that catches 72% of AgentDojo injections catches 15% on tau-bench. False-positive rates on tool outputs, which range from none to over 90%, do transfer between the two agent benchmarks. Where training data is public, the form of the training inputs explains the results. The BIPIA leader was trained on full BIPIA inputs, but having seen InjecAgent's attack strings as short prompts does not help it find them inside tool outputs; the best detector on both agent benchmarks shares no data with any benchmark and was trained on agent-style inputs. Evaluations meant to inform deployment should use the agent's own tool outputs, report detection at a low false-positive rate, and audit what the detector was trained on.

Ethereum's electricity use fell by about 99.95% after the move from proof of work to proof of stake. Service providers still need to report operational energy use, e.g. under the EU Markets in Crypto-Assets Regulation (MiCAR). Existing estimates either apply one typical wattage to every node or start from aggregated monitoring counts. Both ignore attributes that nodes already advertise on the peer-to-peer network: client software, ARM or x86 hardware, hosting location, and validator role. We crawl the consensus and execution layers, assign each peer a wattage from those attributes using published measurements, and estimate the remaining incomplete peers with a Random Forest. On 6,934 peers from two Nebula crawls (19 and 22 June 2026), reachable nodes sum to 415 kW, or 3.63 GWh if that draw were held for a year. The same Lighthouse+Nethermind x86 wattage on every peer yields 431 kW. Observed attributes lower the total by 3.9%, mainly because nodes at Hetzner and other non-AWS clouds draw less than that home-desktop figure. AWS accounts for 15.6% of watts from 12.2% of peers, and validator-flagged nodes for 31.4% of watts from 25.7% of peers. The 415 kW snapshot is about 46% of the Cambridge Centre for Alternative Finance (CCAF) estimate of about 0.90 MW. Both use about 60 W per node, so the gap is mostly how many nodes each estimate includes. Rules cover 3,110 peers and the forest the other 3,824. On held-out labeled peers with client, architecture, and OS hidden, the forest's mean absolute error against the rule wattage is 4.3 W. Twenty-four-hour measurements on a gaming desktop differ from the predictions. After subtracting a 33 W idle graphics card that Ethereum clients do not need, both differences fall to about 19%.

Large language model-based agents are increasingly deployed to perform domain-specific tasks by interacting with enterprise knowledge, tools, and external services. Existing runtime guardrails primarily target prompt injection and other attack-specific behaviors under a black-box threat model, but provide limited guarantees that agents operate within their intended functionality. As a result, production agents remain vulnerable to malicious requests and out-of-domain queries that existing defenses often fail to distinguish. We present Persona Guardrail, a production-grade runtime defense framework that enforces explicit functional boundaries for customer-facing agentic AI systems through synchronous input and output validation driven by semantic allowlist and blocklist specifications. We also introduce PAGE (Persona-Aware Guardrail Evaluation), a benchmark for evaluating function-specific guardrails across benign, adversarial, and out-of-domain interactions on both user and agent turns. Compared with a generic LLM-based guardrail, Persona Guardrail improves overall accuracy from 85.7% to 95.9%, increases out-of-domain detection from 57.3% to 93.5%, and reduces the false-approved rate from 25.0% to 4.7%. Currently deployed in production, Persona Guardrail meets its latency budget while sustaining a very low false-block and false-allow rate under realistic production workloads. These results demonstrate that Persona Guardrail provides a practical, scalable, and production-ready foundation for securing agentic AI systems.

Large Language Models (LLMs) increasingly support Uncrewed Aerial Vehicle (UAV) swarm operations such as data collection scheduling, where the model reads structured sensor reports and decides which sensors to visit. An adversary who quietly manipulates those reports can redirect the swarm without modifying the model weights or the UAV. Defenses for this interface have been proposed architecturally but rarely implemented or evaluated. We implement and evaluate defense-in-depth at the perception-reasoning interface of LLM-Centric Agentic UAV Swarms. Five layers check the provenance of a report, whether its values are physically admissible, whether they agree with what swarm geometry and service history predict, whether the resulting schedule starves any sensor, and, when these fail, hand control to a deterministic scheduler that ignores the suspect input. We test each layer against an adversary strong enough to defeat the layer before it. For each of the three input-side layers, we derive in closed form how far a report can be distorted before that layer reacts, fixing each boundary from deployment parameters before any attack data is collected; across thirty matched simulation runs, predicted and measured boundaries agree. Separating attack detection from response is a well-established principle, and we quantify the cost of neglecting this distinction at the perception-reasoning interface. When the system rejects a report, it replaces it with the most recent accepted report. This prevents the adversary from controlling the UAV schedule, but it also increases cumulative cost by 79% and 74% for the two detectors, respectively, compared with the undefended system. The safety check does not detect any attacks, but it nevertheless reduces the attack-induced cost by 37.5%.

As the dominant trading mechanism in decentralized finance, Automated Market Maker has been widely studied in research. However, limited research has been done with the trading fees taken into consideration. In this work, we study how much trading fee Liquidity Providers(LPs) can receive from arbitrage trading when the fee rate approaches zero. We give a closed-form formula for it and our result shows that the trading fees generated by arbitrage trading can fully offset the LVR loss when the price process is continuous. We further extend our conclusions to price processes with jumps and the theoretical analysis shows that jumps are the only cause of LP loss apart from market risks. Our results provide practical guidance for AMM designers as well as LPs.

Hardware reverse engineering supports competitive analysis and hardware assurance by recovering information about an integrated circuit (IC) from its physical implementation. While existing techniques primarily recover the gate-level netlist, which represents logical functionality, they often overlook physical design decisions such as placement, routing, delay insertion, and interconnect optimization. The clock distribution network encapsulates many of these decisions; however, no prior published work has recovered this network from a fabricated IC to infer design intent. We present a layout-aware methodology that recovers and analyzes the clock distribution network by integrating a recovered gate-level netlist with layout information extracted from scanning electron microscope imagery. Our four-phase pipeline recovers clock-tree topology, buffering, gating and switching, interconnect delay, and crosstalk-mitigation measures. We demonstrate our methodology on a commercial 450 nm IC, recovering a global H-tree backbone with local X-tree-like branching, identifying an independent clock tree and global clock gating, quantifying latency, skew, and routing lengths, and confirming the absence of dedicated crosstalk mitigation. Together, these findings let us reason about the designer's intent. To encourage further research and support reproducibility, we release our clock tree recovery algorithm as open source.

Digital watermarking supports source attribution for AI-generated images, but its reliability depends on resistance to removal attacks. Some attacks attempt to remove watermarks by forcing the decoded watermark to differ from the original. However, this can produce an inverted watermark that remains detectable, causing removal to fail, while further attempts to alter the watermark may unnecessarily degrade image quality. To address these limitations, we present LiBRA (Latent In-band Bidirectional Removal Attack), which aims to make watermarks undetectable while preserving image quality. Instead of continually pushing the watermark toward inversion, LiBRA adjusts the image to conceal the watermark without encouraging further changes that could degrade image quality. Some attacks keep pushing decoded bits away from the original watermark, even when further changes preserve detectability and damage image quality. With access to the watermark key and decoder, LiBRA makes bounded changes in a public autoencoder's latent space. Unlike inversion-driven objectives that cannot correct excessive inversion, LiBRA guides average decoding confidence toward random guessing from either direction. This helps avoid an inverted but detectable watermark. Leaving individual bits flexible allows image-quality constraints to favor less damaging changes, while an optional frequency-guided mask limits their location. We verify removal using an exact two-sided binomial test rather than assuming the confidence target guarantees success.

Workspace agents combine large language models with execution harnesses to perform stateful, multi-step tasks that access or modify external resources. Existing benchmarks leave gaps in executable coverage of their runtime security risks, while evolving model capabilities, harnesses, tools, and threats motivate benchmark evolution. We introduce EvoRiskBench, an evolving benchmark organized around the EP-Path-EF framework, which links an initial risk entry point to a one-hop technical effect through an agent-mediated risk path. The framework defines nine entry-point categories and five effect categories; a 20-participant study supports their interpretability and classification consistency on representative cases. Guided by this framework, an automated end-to-end workflow constructs and executes risk cases in isolated environments and independently verifies outcomes using runtime traces and environment states. The benchmark provides a reproducible dataset of 450 adversarial tasks across six scenarios. We evaluate nine model-harness configurations spanning three models (GPT-5.6 Sol, DeepSeek-V4-Pro-0813, and Claude Opus 5) and three harnesses (Claude Code, Codex, and OpenClaw). Our results reveal substantial vulnerabilities across systems. The most vulnerable configuration, Codex with DeepSeek-V4-Pro-0813, reaches a 68.44% attack success rate (ASR), indicating that configuration of workspace agent is insufficient to ensure secure autonomous execution. ASR varies more across models than harnesses, and harness differences depend on the model. The benchmark cases and evaluation platform will be released after completion of artifact safety and reproducibility checks.

Open-weight language models can be downloaded, modified, and deployed beyond their developers' control, limiting the effectiveness of centrally enforced safeguards. Recent work has therefore proposed trigger-tag mechanisms that produce a detectable signal when a model is used under a target condition, such as generating phishing contents. Although these mechanisms borrow from established techniques, their use for conditional misuse detection in open-weight LLMs is relatively new. Therefore, existing research works have not systematically studied the robustness of trigger-tag mechanisms under adversarial attacks. To close this gap, (i)~we formalize trigger-tags and distinguish token-level trigger-tags, which introduce watermark-inspired signals during decoding, from weight-level trigger-tags, which learn backdoor-inspired associations between target conditions and detectable model behavior. Furthermore, (ii)~we introduce \Untag, a unified attack framework that organizes their mechanism-specific attack surfaces into a common taxonomy. We evaluate representative token-level and weight-level trigger-tags using phishing as a case study. We find that while trigger-tags may provide useful evidence in controlled settings, our attacks render the existing trigger-tag mechanisms to be entirely ineffective. Consequently, we argue that these mechanisms should not be treated as robust misuse detectors when attackers can transform outputs or modify open weights.