Papers

Cloud networks increasingly rely on machine learning based Network Intrusion Detection Systems to defend against evolving cyber threats. However, real-world deployments are challenged by limited labeled data, non-stationary traffic, and adaptive adversaries. While semi-supervised learning can alleviate label scarcity, most existing approaches implicitly assume benign and stationary unlabeled traffic, leading to degraded performance in adversarial cloud environments. This paper proposes a robust semi-supervised temporal learning framework for cloud intrusion detection that explicitly addresses adversarial contamination and temporal drift in unlabeled network traffic. Operating on flow-level data, this framework combines supervised learning with consistency regularization, confidence-aware pseudo-labeling, and selective temporal invariance to conservatively exploit unlabeled traffic while suppressing unreliable samples. By leveraging the temporal structure of network flows, the proposed method improves robustness and generalization across heterogeneous cloud environments. Extensive evaluations on publicly available datasets (CIC-IDS2017, CSE-CIC-IDS2018, and UNSW-NB15) under limited-label conditions demonstrate that the proposed framework consistently outperforms state-of-the-art supervised and semi-supervised network intrusion detection systems in detection performance, label efficiency, and resilience to adversarial and non-stationary traffic.

Passwords still remain a dominant authentication method, yet their security is routinely subverted by predictable user choices and large-scale credential leaks. Automated password guessing is a key tool for stress-testing password policies and modeling attacker behavior. This paper applies LLM-driven evolutionary computation to automatically optimize prompts for the LLM password guessing framework. Using OpenEvolve, an open-source system combining MAP-Elites quality-diversity search with an island population model we evolve prompts that maximize cracking rate on a RockYou-derived test set. We evaluate three configurations: a local setup with Qwen3 8B, a single compact cloud model Gemini-2.5 Flash, and a two-model ensemble of frontier LLMs. The approach raises the cracking rates from 2.02% to 8.48%. Character distribution analysis further confirms how evolved prompts produce statistically more realistic passwords. Automated prompt evolution is a low-barrier yet effective way to strengthen LLM-based password auditing and underlining how attack pipelines show tendency via automated improvements.

Prompt injection has emerged as a critical security threat to large language models (LLMs), yet existing studies predominantly focus on single-dimensional attack strategies, such as semantic rewriting or character-level obfuscation, which fail to capture the combined effects of multi-space perturbations in realistic scenarios. In addition, systematic black-box robustness evaluations of recent Chinese LLMs, such as DeepSeek, remain limited. To address these gaps, we propose PromptFuzz-SC, a semantic-character dual-space mutation framework for evaluating LLM robustness against prompt injection. The framework integrates semantic transformations (e.g., paraphrasing and word-order perturbation) with character-level obfuscation (e.g., zero-width insertion and encoding-based mutation), forming a unified and extensible mutation operator library. A hybrid search strategy combining epsilon-greedy exploration and hill-climbing refinement is adopted to efficiently discover high-quality adversarial prompts. We further introduce a unified evaluation protocol based on three metrics: misuse success rate (MSR), Average Queries to Success (AQS), and Stealth. Experimental results on DeepSeek demonstrate that dual-space mutation achieves the strongest overall attack performance among the evaluated strategies, attaining the highest mean MSR (0.189), peak MSR (0.375), and mean Stealth. Compared with semantic-only and character-only mutation, it improves mean MSR by 12.5% and 5.6%, respectively. While not consistently minimizing query cost, the proposed method achieves competitive best-case efficiency and maintains strong imperceptibility, indicating a more favorable balance between attack effectiveness and concealment. These findings highlight the importance of composite mutation strategies for robust red-teaming of LLMs and provide practical insights for the design of multi-layer defense mechanisms.

Specification gaming under Reinforcement Learning (RL) is known to cause LLMs to develop sycophantic, manipulative, or deceptive behavior, yet the conditions under which this occurs remain unclear. We train 11 instruction-tuned LLMs (0.5B-14B) with on-policy RL across 3 environments and find that model size acts as a safety buffer in some environments but enables greater harmful exploitation in others. Controlled ablations trace this reversal to environment-specific features such as role framing and implicit gameability cues. We further show that most safety benchmarks do not predict RL-induced misalignment, except in the case of Sycophancy scores when the exploit relies on inferring the user's preference. Finally, we find that on-policy RL preserves a safety buffer inherent in the model's own generation distribution, one that is bypassed during off-policy settings.

Text-to-image (T2I) diffusion models have achieved remarkable success in image synthesis, but their reliance on large-scale data and open ecosystems introduces serious backdoor security risks. Existing defenses, particularly input-level methods, are more practical for deployment but often rely on observable anomalies that become unreliable under stealthy, semantics-preserving trigger designs. As modern backdoor attacks increasingly embed triggers into natural inputs, these methods degrade substantially, raising a critical question: can more stable, implicit, and trigger-agnostic differences between benign and backdoor inputs be exploited for detection? In this work, we address this challenge from an active probing perspective. We introduce controlled scaling perturbations on cross-attention and uncover a novel phenomenon termed Cross-Attention Scaling Response Divergence (CSRD), where benign and backdoor inputs exhibit systematically different response evolution patterns across denoising steps. Building on this insight, we propose SET, an input-level backdoor detection framework that constructs response-offset features under multi-scale perturbations and learns a compact benign response space from a small set of clean samples. Detection is then performed by measuring deviations from this learned space, without requiring prior knowledge of the attack or access to model training. Extensive experiments demonstrate that SET consistently outperforms existing baselines across diverse attack methods, trigger types, and model settings, with particularly strong gains under stealthy implicit-trigger scenarios. Overall, SET improves AUROC by 9.1% and ACC by 6.5% over the best baseline, highlighting its effectiveness and robustness for practical deployment.

Organisations increasingly outsource privacy-sensitive data transformations to cloud providers, yet no practical mechanism lets the data owner verify that the contracted algorithm was faithfully executed. VeriX-Anon is a multi-layered verification framework for outsourced Target-Driven k-anonymization combining three orthogonal mechanisms: deterministic verification via Merkle-style hashing of an Authenticated Decision Tree, probabilistic verification via Boundary Sentinels and exact-duplicate Twins with cryptographic identifiers, and utility-based verification via Explainable AI fingerprinting that compares SHAP value distributions before and after anonymization using the Wasserstein distance. Across seven cross-domain datasets and four cloud profiles (28 scenarios), against Lazy (drops records), Dumb (fake hash), and Approximate (valid hash) adversaries, VeriX-Anon detects 25 of 28 deviations under a fixed threshold and 27 of 28 once the threshold is calibrated per dataset, with no false alarms. No single layer achieved this alone. The XAI layer was the only mechanism that caught the Approximate adversary, succeeding on six of seven datasets and missing only a high-dimensional case where honest generalization shifts SHAP as much as the attack. Target-Driven anonymization preserved significantly more utility than blind splitting, with mean F1 gaps of +0.058 to +0.362 and Wilcoxon p <= 0.001 on six of seven datasets. Client-side verification completes under one second at one million rows. The threat model covers three empirically evaluated profiles and one theoretical Informed Attacker unable to defeat the cryptographic salt. Sentinel evasion probability ranges from near-zero to 0.82 for the most imbalanced data, which the twin layer offsets in every scenario.

Cryptographic agility is a key prerequisite for maintaining the long-term security of digital communication, particularly in light of the transition to post-quantum cryptography. To systematically assess this capability, Hohm et al. proposed the Crypto Agility Maturity Model (CAMM). In this work, we present the first evaluation of the CAMM against established design principles for maturity models. Our analysis reveals that the CAMM only partially satisfies these principles: its scope and target groups remain ambiguous; acceptance criteria are insufficiently operationalized, limiting verifiability and replicability; and dependency relations exhibit redundancies, cycles, and omissions. Applying the CAMM to a simple real-world scenario further confirmed these issues, as several requirements at higher maturity levels proved inapplicable or unclear. Based on these findings, we propose concrete improvements to the CAMM to enable more consistent and reliable assessments of cryptographic agility.

Autonomous vehicles (AVs) promise efficient, clean and cost-effective transportation systems, but their reliance on sensors, wireless communications, and decision-making systems makes them vulnerable to cyberattacks and physical threats. This chapter presents novel design techniques to strengthen the security and resilience of AVs. We first provide a taxonomy of potential attacks across different architectural layers, from perception and control manipulation to Vehicle-to-Any (V2X) communication exploits and software supply chain compromises. Building on this analysis, we present an AV Resilient architecture that integrates redundancy, diversity, and adaptive reconfiguration strategies, supported by anomaly- and hash-based intrusion detection techniques. Experimental validation on the Quanser QCar platform demonstrates the effectiveness of these methods in detecting depth camera blinding attacks and software tampering of perception modules. The results highlight how fast anomaly detection combined with fallback and backup mechanisms ensures operational continuity, even under adversarial conditions. By linking layered threat modeling with practical defense implementations, this work advances AV resilience strategies for safer and more trustworthy autonomous vehicles.

Classical computability theory tells us that self-modifying code (SMC) on a deterministic universal Turing machine can be simulated by non-SMC code on the same model. That abstraction, however, omits the external timing inputs, concurrency, and microarchitectural state that dominate practical execution on modern processors. We argue that once timing, ordering, and self-introspective effects are treated as observables, a practically faithful non-SMC reproduction of timed SMC becomes detectably expensive on commodity systems. We present a tamper-proofing model that combines introspective and polymorphic SMC, reliable clocks, and runtime timing predicates to bind integrity checks to execution behavior. We distinguish static and dynamic SMC generation, characterize the timing semantics needed to avoid catastrophic pipeline clears, and give x86-64 design primitives for checksum-driven self-patching. We also report timer measurements, performance comparisons, and performance-monitoring counter evidence showing that careful engineering – especially loop unrolling and cross-page modification – substantially reduces the overhead of SMC while preserving its tamper-detection value. The paper concludes with an efficiency analysis, a threat model, and deployment guidance for trusted code executing in untrusted environments.

Safety-aligned large language models (LLMs) are increasingly deployed in real-world pipelines, yet this deployment also enlarges the supply-chain attack surface: adversaries can distribute backdoored checkpoints that behave normally under standard evaluation but jailbreak when a hidden trigger is present. Recent post-hoc weight-editing methods offer an efficient approach to injecting such backdoors by directly modifying model weights to map a trigger to an attacker-specified response. However, existing methods typically optimize a token-level mapping that forces an affirmative prefix (e.g., “Sure”), which does not guarantee sustained harmful output – the model may begin with apparent agreement yet revert to safety-aligned refusal within a few decoding steps. We address this reliability gap by shifting the backdoor objective from surface tokens to internal representations. We extract a steering vector that captures the difference between compliant and refusal behaviors, and compile it into a persistent weight modification that activates only when the trigger is present. To preserve stealthiness and benign utility, we impose a null-space constraint so that the injected edit remains dormant on clean inputs. The method is efficient, requiring only a small set of examples and admitting a closed-form solution. Across multiple safety-aligned LLMs and jailbreak benchmarks, our method achieves high triggered attack success while maintaining non-triggered safety and general utility.

Training models on a carefully chosen portion of data rather than the full dataset is now a standard preprocess for modern ML. From vision coreset selection to large-scale filtering in language models, it enables scalability with minimal utility loss. A common intuition is that training on fewer samples should also reduce privacy risks. In this paper, we challenge this assumption. We show that subset training is not privacy free: the very choices of which data are included or excluded can introduce new privacy surface and leak more sensitive information. Such information can be captured by adversaries either through side-channel metadata from the subset selection process or via the outputs of the target model. To systematically study this phenomenon, we propose CoLA (Choice Leakage Attack), a unified framework for analyzing privacy leakage in subset selection. In CoLA, depending on the adversary's knowledge of the side-channel information, we define two practical attack scenarios: Subset-aware Side-channel Attacks and Black-box Attacks. Under both scenarios, we investigate two privacy surfaces unique to subset training: (1) Training-membership MIA (TM-MIA), which concerns only the privacy of training data membership, and (2) Selection-participation MIA (SP-MIA), which concerns the privacy of all samples that participated in the subset selection process. Notably, SP-MIA enlarges the notion of membership from model training to the entire data-model supply chain. Experiments on vision and language models show that existing threat models underestimate subset-training privacy risks: the expanded privacy surface leaks both training and selection membership, extending risks from individual models to the broader ML ecosystem.

As cross-chain interoperability advances, decentralized finance (DeFi) protocols enable illicit funds to be reorganized into uniform liquid assets that flow throughout the cryptocurrency market. Such operations can bypass monitoring targeted at individual blockchains and thereby weaken current regulatory frameworks. Motivated by these, we introduce UniDetect, a multi-chain cryptocurrency fraud account detection method based on large language models (LLMs). Specifically, we use domain knowledge to guide the LLM to generate general transaction summary texts applicable to heterogeneous blockchain accounts, which serve as evidence for fraud account detection. Furthermore, we introduce a two-stage alternating training strategy to continuously and dynamically enhance the multimodal joint reasoning for detecting fraudulent accounts based on both the textual evidence and the transaction graph patterns. Experiments on multiple blockchains show that UniDetect outperforms existing methods 5.57% to 7.58% in Kolmogorov-Smirnov (KS). For cross-chain zero-shot detection, UniDetect identifies over 94.58% of fraudulent accounts. It also generalizes well to non-blockchain data, delivering a 6.06% improvement in F1 over existing methods. The dataset and source code are available at https://github.com/msy0513/UniDetect.

Web agents powered by vision-language models (VLMs) enable autonomous interaction with web environments by perceiving and acting on both visual and textual webpage content to accomplish user-specified tasks. However, they are highly vulnerable to prompt injection attacks, where adversarial instructions embedded in HTML or rendered screenshots can manipulate agent behavior and lead to harmful outcomes such as information leakage. Existing defenses, including system prompt defenses and direct fine-tuning of agents, have shown limited effectiveness. To address this issue, we propose a defense framework in which a web agent operates in parallel with a dedicated guard agent, decoupling prompt injection detection from the agent's own reasoning. Building on this framework, we introduce WebAgentGuard, a reasoning-driven, multimodal guard model for prompt injection detection. We construct a synthetic multimodal dataset using GPT-5 spanning 164 topics and 230 visual and UI design styles, and train the model via reasoning-intensive supervised fine-tuning followed by reinforcement learning. Experiments across multiple benchmarks show that WebAgentGuard consistently outperforms strong baselines while preserving agent utility, without introducing additional latency.

SpanKey is a lightweight way to gate inference without encrypting weights or chasing leaderboard accuracy on gated inference. The idea is to condition activations on secret keys. A basis matrix $B$ defines a low-dimensional key subspace $Span(B)$; during training we sample coefficients $\alpha$ and form keys $k=\alpha^\top B$, then inject them into intermediate activations with additive or multiplicative maps and strength $\gamma$. Valid keys lie in $Span(B)$; invalid keys are sampled outside that subspace. We make three points. (i) Mechanism: subspace key injection and a multi-layer design space. (ii) Failure mode: key absorption, together with two analytical results (a Beta-energy split and margin-tail diagnostics), explains weak baseline separation in energy and margin terms – these are not a security theorem. iii) Deny losses and experiments: Modes A–C and extensions, with CIFAR-10 ResNet-18 runs and MNIST ablations for Mode B. We summarize setup and first-order analysis, injectors, absorption, deny losses and ablations, a threat discussion that does not promise cryptography, and closing remarks on scale. Code: https://github.com/mindmemory-ai/dksc

Large Language Models (LLMs) are increasingly deployed across diverse domains, yet their vulnerability to jailbreak attacks, where adversarial inputs bypass safety mechanisms to elicit harmful outputs, poses significant security risks. While prior work has primarily focused on prompt injection attacks, these approaches often require resource-intensive prompt engineering and overlook other critical components, such as chat templates. This paper introduces TEMPLATEFUZZ, a fine-grained fuzzing framework that systematically exposes vulnerabilities in chat templates, a critical yet underexplored attack surface in LLMs. Specifically, TEMPLATEFUZZ (1) designs a series of element-level mutation rules to generate diverse chat template variants, (2) proposes a heuristic search strategy to guide the chat template generation toward the direction of amplifying the attack success rate (ASR) while preserving model accuracy, and (3) integrates an active learning-based strategy to derive a lightweight rule-based oracle for accurate and efficient jailbreak evaluation. Evaluated on twelve open-source LLMs across multiple attack scenarios, TEMPLATEFUZZ achieves an average ASR of 98.2% with only 1.1% accuracy degradation, outperforming state-of-the-art methods by 9.1%-47.9% in ASR and 8.4% in accuracy degradation. Moreover, even on five industry-leading commercial LLMs where chat templates cannot be specified, TEMPLATEFUZZ attains a 90% average ASR via chat template-based prompt injection attacks.

Cyber Threat Intelligence (CTI) reports contain Indicators of Compromise (IOCs) that are critical for security operations. To operationalize these IOCs across heterogeneous logs, analysts often convert them into regular expressions (regexes) for tasks such as digital forensics, log parsing, and SIEM rule creation. However, regex construction is still largely manual, requiring analysts to extract IOCs from CTI reports and transform them into syntactically valid and semantically precise patterns. This process is slow, error-prone, and increasingly impractical as CTI volumes grow. Although recent studies have applied Large Language Models (LLMs) to IOC extraction, they typically output plain strings rather than regexes, limiting practical deployment. Plain IOCs cannot effectively capture variations in system context, log format, or attacker behavior. To address this gap, we propose IOCRegex-gen, a fully automated LLM-based regex generation system that converts IOCs into regexes. The system introduces two key innovations: (i) a group-aware mechanism that identifies which IOC segments should be represented as capture or non-capture groups, and (ii) an iterative reasoning and multi-stage validation pipeline to ensure syntactic validity and semantic correctness. Experiments on over 3,000 real CTI reports and 2,400 ground-truth strings from the MITRE ATT&CK Evaluation framework show that IOCRegex-gen achieves an average hit rate of 99.1% and a false-positive rate of only 0.8%, demonstrating its effectiveness for large-scale CTI processing and automated regex generation.

The widespread use of Large Language Models (LLMs) in text generation has raised increasing concerns about intellectual property disputes. Watermarking techniques, which embed meta information into AI-generated content (AIGC), have the potential to serve as judicial evidence. However, existing methods rely on statistical signals in token distributions, leading to inherently probabilistic detection and reduced reliability, especially in multi-bit encoding (e.g., timestamps). Moreover, such methods introduce detectable statistical patterns, making them vulnerable to forgery attacks and enabling model providers to fabricate arbitrary watermarks. To address these issues, we propose the concept of trustworthy watermark, which achieves reliable recovery with 100% identification accuracy while resisting both user-side statistical attacks and provider-side forgery. We focus on trustworthy time watermarking for use as judicial evidence. Our framework integrates cryptographic techniques and encodes time information into time-dependent secret keys under regulatory supervision, preventing arbitrary timestamp fabrication. The watermark payload is decoupled from time and generated as a random, non-stored bit sequence for each instance, eliminating statistical patterns. To ensure verifiability, we design a two-stage encoding mechanism, which, combined with error-correcting codes, enables reliable recovery of generation time with theoretically perfect accuracy. Both theoretical analysis and experiments demonstrate that our framework satisfies the reliability requirements for judicial evidence and offers a practical solution for future AIGC-related intellectual property disputes.

Industrial control systems operate in dynamic environments where traffic distributions vary across scenarios, labeled samples are limited, and unknown attacks frequently emerge, posing significant challenges to cross-domain intrusion detection. To address this issue, this paper proposes a clustering-enhanced domain adaptation method for industrial control traffic. The framework contains two key components. First, a feature-based transfer learning module projects source and target domains into a shared latent subspace through spectral-transform-based feature alignment and iteratively reduces distribution discrepancies, enabling accurate cross-domain detection. Second, a clustering enhancement strategy combines K-Medoids clustering with PCA-based dimensionality reduction to improve cross-domain correlation estimation and reduce performance degradation caused by manual parameter tuning. Experimental results show that the proposed method significantly improves unknown attack detection. Compared with five baseline models, it increases detection accuracy by up to 49%, achieves larger gains in F-score, and demonstrates stronger stability. Moreover, the clustering enhancement strategy further boosts detection accuracy by up to 26% on representative tasks. These results suggest that the proposed method effectively alleviates data scarcity and domain shift, providing a practical solution for robust cross-domain intrusion detection in dynamic industrial environments.

Protecting sensitive visual content from unauthorized redistribution is a growing challenge for privacy focused mobile applications, including dating platforms. Screenshot prevention mechanisms, rely on server side monitoring or are limited to digital screenshot detection, are commonly deployed to stop forwarding sensitive images. However, an adversary uses another smartphone to take a photo of the mobile screen, in this scenario the existing solutions offer no protection against psychically screen recapture attacks. Since the attack happens in the physical plane rather than on a digital plane and shows a void or hole in the existing solutions, we name this the Screen Recaptured Analog Hole Attack (S RAHA). Such physically recaptured images bypass digital safeguards and can be freely forwarded, creating substantial privacy, personal safety, and forensic risks. We present a low computational secure by design on device framework that aims to detect and prevent the forwarding of recaptured images directly to the users device. The proposed system integrates a deep learning assisted recapture detection model capable of distinguishing original digital content from camera to screen captures under diverse environmental conditions, together with an on device enforcement mechanism that automatically blocks the sharing of suspected recaptured images between applications. We also introduce the concept of an invisible metadata identifier (IMI) that can be embedded into protected images to enable forensic traceability of potential leakage paths. Although the IMI component is explored at a conceptual and feasibility level rather than fully implemented, it demonstrates a promising direction for integrating lightweight, invisible identifiers into client side security architectures.

LLM-based agents can execute actions that are syntactically valid, user-sanctioned, and semantically appropriate, yet still violate organizational policy because the facts needed for correct policy judgment are hidden at decision time. We call this failure mode policy-invisible violations: cases in which compliance depends on entity attributes, contextual state, or session history absent from the agent's visible context. We present PhantomPolicy, a benchmark spanning eight violation categories with balanced violation and safe-control cases, in which all tool responses contain clean business data without policy metadata. We manually review all 600 model traces produced by five frontier models and evaluate them using human-reviewed trace labels. Manual review changes 32 labels (5.3%) relative to the original case-level annotations, confirming the need for trace-level human review. To demonstrate what world-state-grounded enforcement can achieve under favorable conditions, we introduce Sentinel, an enforcement framework based on counterfactual graph simulation. Sentinel treats every agent action as a proposed mutation to an organizational knowledge graph, performs speculative execution to materialize the post-action world state, and verifies graph-structural invariants to decide Allow/Block/Clarify. Against human-reviewed trace labels, Sentinel substantially outperforms a content-only DLP baseline (68.8% vs. 93.0% accuracy) while maintaining high precision, though it still leaves room for improvement on certain violation categories. These results demonstrate what becomes achievable once policy-relevant world state is made available to the enforcement layer.

We present COBALT-TLA, a neuro-symbolic verification loop that pairs an LLM with TLC, the TLA+ model checker, in an automated REPL. The LLM generates bounded TLA+ specifications; TLC acts as a semantic oracle; structured error traces are parsed and injected back into the model's context to drive convergence. We evaluate the system against three cross-chain bridge targets, including a faithful model of the Nomad $190M exploit. COBALT-TLA reaches a verified BUG_FOUND state in at most 2 iterations on all targets, with TLC execution consistently below 0.30 seconds. Notably, the system autonomously discovers an unprompted vulnerability class – the Optimistic Relay Attack – not present in the human-written baseline specification. We argue that deterministic prover feedback is sufficient to neutralize LLM hallucination in formal methods, transforming zero-shot code generation into a convergent proof-finding strategy.

The applications of Generative Artificial Intelligence (GenAI) and their intersections with data-driven fields, such as healthcare, finance, transportation, and information security, have led to significant improvements in service efficiency and low latency. However, this synergy raises serious concerns regarding the security of large language models (LLMs) and their potential impact on the privacy of companies and users' data. Many technology companies that incorporate LLMs in their services with a certain level of command and control bear a risk of data exposure and secret divulgence caused by insecure LLM pipelines, making them vulnerable to multiple attacks such as data poisoning, prompt injection, and model theft. Although several security techniques (input/output sanitization, decentralized learning, access control management, and encryption) were implemented to reduce this risk, there is still an imminent risk of quantum computing attacks, which are expected to break existing encryption algorithms, hence, retrieving secret keys, encrypted sensitive data, and decrypting encrypted models. In this extensive work, we integrate the Post-Quantum Cryptography (PQC) based Lattice-based Homomorphic Encryption (HE) main functions in the LLM's inference pipeline to secure some of its layers against data privacy attacks. We modify the inference pipeline of the transformer architecture for the LLAMA-3 model while injecting the main homomorphic encryption operations provided by the concrete-ml library. We demonstrate high text generation accuracies (up to 98%) with reasonable latencies (237 ms) on an i9 CPU, reaching up to 80 tokens per second, which proves the feasibility and validity of our work while running a FHE-secured LLAMA-3 inference model. Further experiments and analysis are discussed to justify models' text generation latencies and behaviours.

This paper investigates covert wireless communication with a Fusion Center (FC) that aggregates raw energy measurements from multiple Wardens via soft fusion. Extending our prior work on power-threshold randomization, we consider a stronger adversarial model in which FC randomizes both the number of active Wardens W and the detection threshold t, while Alice and a friendly Jammer jointly randomize their transmit powers under an outage constraint at Bob. We derive a closed-form expression for FC's optimal soft-fusion threshold and show that it is independent of the number of active Wardens. Thus, strategic uncertainty in the sensing infrastructure provides no meaningful detection advantage for FC under soft fusion. We further establish a robustness theorem showing that, even under arbitrary FC randomization over (W,t), Alice and Jammer can maintain outage-feasible communication at Bob while preserving covertness with high probability, provided their power ranges are sufficiently large. This reveals a structural limitation of soft fusion. A game-theoretic formulation characterizes the Nash equilibrium mixed strategies of both sides, accounting for deployment costs and detection-pressure parameters. Analytical and numerical results show that: 1) soft fusion is largely insensitive to the number of Wardens; 2) even semi-strategic finite-support geometric randomization of W performs comparably to the full game-theoretic equilibrium; and 3) the covertness-reliability tradeoff remains nearly invariant across a wide range of FC deployment costs and strategy parameters. These findings exemplify a Red Queen effect, in which FC incurs increasing operational costs for only marginal gains in detection performance, and highlight the need for alternative detection architectures.

Covert wireless communication aims to establish a reliable link while hiding the transmission from an adversary. In wireless settings, uncertainty plays a central role in this tradeoff: it can help mask the signal from a warden, but it also complicates robust system design. This raises a basic question: under bounded uncertainty, are reliability and covertness governed by the same adverse conditions? If not, robust covert design cannot be reduced to a single worst-case environment. In this paper, we study this question in a covert wireless model with quasi-static fading, outage-based reliability at Bob and radiometric detection at Willie. Uncertainty is represented through bounded intervals for Bob's average channel strength and Willie's noise power. To obtain a tractable characterization, we adopt a conditional large-N midpoint-threshold surrogate for Willie's detector, parameterized by a Willie-side fading realization. Within this framework, we show that the reliability constraint is governed by Bob's smallest admissible channel parameter, whereas the covertness constraint is governed by Willie's smallest admissible noise level. This establishes a conflict-aware robust-design principle: the adverse realizations for reliability and covertness differ. Based on this result, we derive closed-form expressions for the robustly feasible transmit power and the corresponding robust optimal rate. Numerical results show that bounded uncertainty contracts the feasible region, monotonically reduces the robust optimal rate, and can cause substantial loss relative to the nominal design. Monte Carlo results further show that the conditional surrogate closely tracks the midpoint-threshold radiometer in the intended low-effective-SNR regime. Overall, the paper shows that even in a streamlined wireless setting, robust covert design requires different adverse-case reasoning for reliability and covertness.

Covert quantum communication is usually analyzed under idealized assumptions that channel parameters, such as transmissivity and background noise, are perfectly known and constant. In realistic optical links, including satellite, fiber, and free-space systems, these parameters vary because of environmental fluctuations, calibration noise, and estimation errors. We study covert quantum communication over compound quantum optical channels with bounded uncertainty in both transmissivity and thermal noise, and derive guarantees that hold for all admissible channel realizations. We develop a robust framework for certifying both covertness and reliability under uncertainty. A central finding is that robustness cannot be obtained by simply inserting worst-case parameter values into known-channel bounds: the channel realizations that are most adverse for covertness and reliability generally occur at different corners of the uncertainty set. This creates a fundamental trade-off in secure system design. We derive a closed-form lower bound on the worst-case guaranteed number of covert qubits that can be transmitted reliably, identify a sharp feasibility boundary beyond which the guaranteed payload drops to zero, and quantify the security penalty caused by uncertainty. We validate the covertness term with QuTiP simulations of a four-mode bosonic model and combine it with an analytical reliability bound to evaluate the robust payload. Our results move covert quantum communication from nominal perfect-knowledge analysis to certified worst-case operation under uncertainty.

Coding agents and LLM-powered applications routinely send potentially sensitive content to cloud LLM APIs where it may be logged, retained, used for training, or subpoenaed. Existing privacy tooling focuses on network-level encryption and organization-level DLP, neither of which addresses the content of prompts themselves. We present a systematic empirical evaluation of eight techniques for privacy-preserving LLM requests: (A) local-only inference, (B) redaction with placeholder restoration, (C) semantic rephrasing, (D) Trusted Execution Environment hosted inference, (E) split inference, (F) fully homomorphic encryption, (G) secret sharing via multi-party computation, and (H) differential-privacy noise. We implement all eight (or a tractable research-stage subset where deployment is not yet feasible) in an open-source shim compatible with MCP and any OpenAI-compatible API. We evaluate the four practical options (A, B, C, H) and their combinations across four workload classes using a ground-truth-labelled leak benchmark of 1,300 samples with 4,014 annotations. Our headline finding is that no single technique dominates: the combination A+B+C (route locally when possible, redact and rephrase the rest) achieves 0.6% combined leak on PII and 31.3% on proprietary code, with zero exact leaks on PII across 500 samples. We present a decision rule that selects the appropriate option(s) from a threat-model budget and workload characterisation. Code, benchmarks, and evaluation harness are released at https://github.com/jayluxferro/llm-redactor.

A recent and exciting thread of work focuses on developing methods for watermarking the output of large language models (LLMs). We focus on provably undetectable watermarking-that is, schemes that do not alter the output distribution of the LLM, yet enable embedding a watermark in the output that identifies the output as having been generated by the particular LLM. Furthermore, the watermark should be hard to remove by an adversary that may potentially edit, insert, or delete tokens from the watermarked output. Indeed, recent work (Christ et al. [COLT'24], Christ et al. [CRYPTO'24], Golowich et al. [NeuroIPS'24]) shows how to develop such schemes that are robust against a constant fraction of substitutions, or even against a constant fraction of arbitrary edits. These works, however, make strong assumptions on the entropy present in the output of the LLM. Most notably, they all require constant entropy rate-that is, a constant fraction of the tokens in a sufficiently long substring of the output need to have empirical entropy at least O(log |T|), where T is the alphabet of tokens, and Golowich et al. additionally require T to be larger than the security parameter. In this work, we consider whether we can also watermark the outputs of LLMs when the per-token entropy is just a constant, discarding the dependence on the alphabet size or security parameter. In this regime, we construct: - A watermarking scheme robust against random substitutions (assuming subexponential LPN, as in Christ et al. [CRYPTO'24]) - A watermarking scheme robust against random substitutions and random deletions, given either the additional heuristic assumption that the output of the LLM only introduces random errors (analogous to the assumption made by Christ et al. [CRYPTO'24]) or a construction of a pseudorandom error-correcting code robust to adversarial substitutions and random deletions.

We present SIR-Bench, a benchmark of 794 test cases for evaluating autonomous security incident response agents that distinguishes genuine forensic investigation from alert parroting. Derived from 129 anonymized incident patterns with expert-validated ground truth, SIR-Bench measures not only whether agents reach correct triage decisions, but whether they discover novel evidence through active investigation. To construct SIR-Bench, we develop Once Upon A Threat (OUAT), a framework that replays real incident patterns in controlled cloud environments, producing authentic telemetry with measurable investigation outcomes. Our evaluation methodology introduces three complementary metrics: triage accuracy (M1), novel finding discovery (M2), and tool usage appropriateness (M3), assessed through an adversarial LLM-as-Judge that inverts the burden of proof – requiring concrete forensic evidence to credit investigations. Evaluating our SIR agent on the benchmark demonstrates 97.1% true positive (TP) detection, 73.4% false positive (FP) rejection, and 5.67 novel key findings per case, establishing a baseline against which future investigation agents can be measured.

While recent LLM-based agents can identify many candidate bugs in source code, their reports remain static hypotheses that require manual validation, limiting the practicality of automated bug detection. We frame this challenge as a test generation task: given a candidate report, synthesizing an executable proof-of-concept (PoC) - such as a script, command sequence, or crafted input - to trigger the suspected defect. Automated PoC generation can act as a scalable validation oracle, enabling end-to-end autonomous bug detection by providing concrete execution evidence. However, naive LLM agents are unreliable validators: they are biased toward "success" and may reward-hack by producing plausible but non-functional PoCs or even hallucinated traces. To address this, we present ANYPoC, a general multi-agent framework that (1) analyzes and fact-checks a candidate bug report, (2) iteratively synthesizes and executes a PoC while collecting execution traces, and (3) independently re-executes and scrutinizes the PoC to mitigate hallucination and reward hacking. In addition, ANYPoC also continuously extracts and evolves a PoC knowledge base to handle heterogeneous tasks. ANYPoC operates on candidate bug reports regardless of their source and can be paired with different bug reporters. To demonstrate practicality and generality, we apply ANYPoC, together with a simple agentic bug reporter, on 12 large-scale, critical software systems, including Firefox, Chromium, LLVM, OpenSSL, SQLite, FFmpeg, and Redis. Compared to the state-of-the-art coding agents, e.g., Claude Code and Codex, ANYPoC produces 37% more valid PoCs for true-positive bug reports and rejects 9.7x more false-positive bug reports. ANYPoC also enables the discovery of 121 new bugs from over two thousand noisy bug reports, with 108 confirmed by developers and 92 fixed. 46 PoCs have also been adopted as official regression tests.

Time-series forecasting aims to predict future values by modeling temporal dependencies in historical observations. It is a critical component of many real-world systems, where accurate forecasts improve operational efficiency and help mitigate uncertainty and risk. More recently, machine learning (ML), and especially deep learning (DL)-based models, have gained widespread adoption for time-series forecasting, but they remain vulnerable to adversarial attacks. However, many state-of-the-art attack methods are not directly applicable in time-series settings, where storing complete historical data or performing attacks at every time step is often impractical. This paper proposes an adversarial attack framework for time-series forecasting under an online bounded-buffer setting, leveraging an informed and selective attack strategy. By selectively targeting time steps where the model exhibits high confidence and the expected prediction error is maximal, our framework produces fewer but substantially more effective attacks. Experiments show that our framework can increase the prediction error up to 2.42x, while performing attacks in fewer than 10% of time steps.