arXiv (research papers)

Recent stories from arXiv (research papers).

arXiv (research papers) Pedram MohajerAnsari Clemson University Security

User Perceptions of Deceptive In-Vehicle Infotainment User Interfaces

Modern vehicles collect extensive data about drivers, passengers, and vehicle use, with privacy and consent decisions increasingly mediated through infotainment software. Although dark patterns have been widely studied in web, mobile, and e-commerce interfaces, little is known about how they appear in automotive infotainment systems or how users perceive them. We study this problem through Android Automotive OS (AAOS), conducting a three-phase empirical analysis. First, 26…

Read at arXiv (research papers)

arXiv (research papers) Joshua D Guttman Security

Mutable State in Enrich-by-Need Security Protocol Analysis

Security protocol analysis has emphasized message-passing, as this is how different principals interact in authentication, key exchange, and similar functionalities. However, in many protocols a principal also maintains mutable state that limits what actions its local sessions may take; this cross-session state is altered by other local sessions. In this paper we show how to represent this cross-session mutable state in the strand space approach to symbolic protocol…

Read at arXiv (research papers)

arXiv (research papers) William Fleshman Security

Living off the Local Model: Abusing Browser-Embedded Language Models for Post-Exploitation

Browsers now ship local language models to ordinary endpoints. Google includes Gemini Nano in Chrome to enable productivity features such as summarization, with inference running entirely on the device. These browser-embedded models provide new living-off-the-land opportunities for attackers, each instance a potential post-exploitation engine already installed, signed by the vendor, and invisible to network monitoring. The Gemini model is multimodal, capable of processing…

Read at arXiv (research papers)

arXiv (research papers) Syed Ahzaz Shah Security

PriTabICL: Differentially Private Tabular In-Context Learning with Task-Aligned Cluster-Wise Synthesis

Tabular in-context learning (TabICL) enables large language models (LLMs) to predict target records using a handful of labeled demonstrations in their prompts. However, these demonstrations often contain sensitive personal information. Differential privacy (DP) offers principled protection, but noise-based methods introduce substantial noise, distorting the feature-label relationships essential for accurate prediction, especially as data dimensionality grows. In this work…

Read at arXiv (research papers)

arXiv (research papers) Matthew D. Laws Security

Attacks and Defenses in Decentralized Multi-Agent Systems with Hop-by-Hop Task Decomposition and Delegation

Decentralized multi-agent LLM systems enable independently administered agents, each with its own operator, model, memory, and configuration, to collaborate on complex tasks without centralized coordination. Agents autonomously decompose tasks, execute subtasks and delegate remaining work to peers, enabling scalability and eliminating single points of failure. However, decentralization also introduces significant security challenges, as no single entity has complete…

Read at arXiv (research papers)

arXiv (research papers) Yipeng Wang Security

"I Don't Think It's a Scam": A First Look at Family Intervention Barriers in Elder Abuse Discussions on Reddit and Baidu Tieba

When older adults trust scammers despite warnings, families face obstacles to intervention. Complainants aged 60 and older reported \$4.9 billion in losses in 2024 to the U.S. Internet Crime Complaint Center. To investigate family intervention barriers to elder scams, we collected 2M Reddit posts from May 2021 to June 2026 and 2B Baidu Tieba posts from 2007 to 2026. The final corpora contain 13,082 Reddit posts and 3,933 Baidu Tieba posts, labeled using platform-specific…

Read at arXiv (research papers)

arXiv (research papers) Amr Elshazly Security

LLM-Guided Cyber-Physical Attack Path Identification in Smart Grids

Identifying high-impact attack paths in cyber-physical power systems is critical for improving grid resilience. Existing methods typically rely on either structural topology analysis or simulation-based cascading-failure models. Topology-based approaches capture connectivity but lack adversarial reasoning, while simulation-driven methods are primarily suited for offline analysis and are less effective for rapid evaluation of evolving attack scenarios. As a result…

Read at arXiv (research papers)

arXiv (research papers) Omkar Shrikanth Security

Stealing Exact Training Data from Other Clients during Federated Learning

Malicious adversaries can employ gradient inversion attacks to reverse-engineer private data using the model updates of a round in Federated Learning. While there have been developments in gradient inversion for text, exact text recovery by a malicious client has so far remained infeasible, as FedAvg entangles the gradient contributions of many clients and obscures the signal of any individual sentence. We introduce a novel attack pipeline that unwinds this entanglement…

Read at arXiv (research papers)

arXiv (research papers) Hemen Taukoor Security

The Limit Is in the Input, Not the Model: State-Dependent Command Safety in Natural-Language Interfaces to Industrial Controllers

Language models judging commands to industrial controllers may not receive machine state, the equipment's current condition. The same command can require opposite safety decisions in different states, so an incorrect answer may reflect missing information rather than poor reasoning. We propose an answerability audit: does the model receive the information needed to decide? The audit records the model's inputs and pairs the same command with two states requiring opposite…

Read at arXiv (research papers)

arXiv (research papers) Gonçalo Fernandes Security

Gradient-Based Defensive Poisoning Against System Identification Attacks on Cyber-Physical Systems

AI-driven attacks learn the models of cyber-physical systems (CPS) by passively observing controller and plant signals, then use this knowledge to act covertly. Because covertness depends on the accuracy of the identified model, degrading that accuracy is a viable countermeasure. This work applies Adversarial Machine Learning as a defense, injecting small perturbations into the measurements the attacker observes to degrade an artificial neural network (ANN)-based system…

Read at arXiv (research papers)

arXiv (research papers) Eric Tacl Security

ChainIT Transaction Truth: Stablecoins, Web3 Wallets, and Multi-Rail Commerce. A Zero-Trust Control Plane for Verified Identity, Organizational Authority, Runtime Compliance, Secure Wallet Execution, and Immutable Settlement Evidence

Stablecoins can move value continuously across programmable networks, and Web3 wallets can authorize those transfers with cryptographic signatures, but neither capability, standing alone, establishes a complete commercial transaction. A blockchain may prove that a transaction satisfied protocol rules, and a wallet may prove that sufficient signing authority existed; neither proves who initiated the transaction, for which legal entity, under what current authority, against…

Read at arXiv (research papers)

arXiv (research papers) Haochen Sun Security

Robust, Lightweight Locally Differentially Private Histogram Estimation and the Role of Dropout Attacks

Local differential privacy (LDP) is vulnerable to output manipulation by malicious clients. Zero-knowledge proofs can enforce correct report computation, but clients may still selectively drop out after inspecting their outputs. Prior work combines zero-knowledge proofs with homomorphic encryption to hide reports from clients and prevent selective dropout, but incurs substantial computational cost. We instead develop a three-message lightweight verification framework for…

Read at arXiv (research papers)

arXiv (research papers) Zeta Avarikioti Security

Playing the SMR Game with Existential Honesty

State machine replication protocols are usually proven secure in the honest-Byzantine setting, which requires an honest supermajority of validators. Deployed systems offer no such guarantee: validators participate for profit and may deviate from the protocol whenever doing so increases their gains. In our work, every validator is rational or Byzantine, and none is honest. Rational validators may collude with each other and with the Byzantine ones, who may also bribe them to…

Read at arXiv (research papers)

arXiv (research papers) Yi Yang Security

SCR: Detecting Backdoor-Poisoned Samples through Clean-Reference Residuals

Deep neural networks are vulnerable to backdoor attacks, in which adversaries manipulate model predictions by injecting poisoned samples into training data. Detecting these samples remains challenging because legitimate semantic variation can obscure poisoning-related signals in learned representations. Consequently, class-level statistics used by existing detectors may flag benign feature deviations while overlooking poisoned samples that blend into the class distribution…

Read at arXiv (research papers)

arXiv (research papers) Sean Lim Security

Probing agent overeagerness in realistic settings with continuation audits

Frontier coding agents increasingly exhibit overeagerness, taking surprising, sometimes reckless actions beyond a user's mandate --- reported in recent system cards from frontier model developers. These cases are surfaced through internal deployment monitoring, which requires private traffic and is out of reach for open-weight models and independent evaluators. Automated alignment audits are the natural alternative, but capable models increasingly recognise the synthetic…

Read at arXiv (research papers)

arXiv (research papers) Yun Bian Security

Mitigating fix reversal: Counterfactual constraints for function-level vulnerability detection

Function-level vulnerability detectors often exhibit fix reversal, assigning a higher vulnerability probability to a fixed version than to its vulnerable counterpart. Fixes usually lengthen functions, and larger length increases are associated with more severe reversal. Code-transformation experiments indicate that fix reversal mainly stems from a length shortcut: vulnerability probability increases with function length when content is fixed. We propose CFVul to address this…

Read at arXiv (research papers)

arXiv (research papers) Mohammed A. Alharbi Security

Drift-Aware Explainable Prioritization for Windows PE Malware Forensic Triage with Semantic Consistency Checks

Forensic investigators must decide which suspicious files to examine first. Machine-learning (ML) and deep-learning (DL) malware detectors usually return a bare score that neither explains why a file deserves attention nor signals when changing data make the model unreliable. We present a framework for Windows Portable Executable (PE) malware forensic triage that works only on pre-extracted static features and never executes malware. A calibrated maliciousness score sets…

Read at arXiv (research papers)

arXiv (research papers) Lipeng He Security

PIPETTE: Prompt Injection Purple-Teaming for Robust LLM Agents

Large language model (LLM) agents routinely interact with untrusted external data, exposing them to indirect prompt injection (IPI) attacks that hijack agents into executing attacker-chosen tasks. To uncover these vulnerabilities at scale, automated red-teaming methods often leverage reinforcement learning (RL) to generate attacks that achieve high attack success rates (ASR). Model developers use the attacks discovered during internal red-teaming to harden agents through…

Read at arXiv (research papers)

arXiv (research papers) Stavros Damianakis Security

AutoPI: Taxonomy-Guided Adaptive Red Teaming for Prompt Injection in LLM Agents

Prompt injection is the dominant unsolved security risk for LLM agents, and the community now agrees that robustness must be measured against adaptive attackers rather than fixed benchmark injections. Many adaptive red teamers exist, and most are built to maximize attack success rate (ASR). Yet ASR counts how often a target breaks, not in how many different ways. A red teamer can achieve a high ASR by reusing a few techniques that work, leaving the others untested and absent…

Read at arXiv (research papers)

arXiv (research papers) Chengzhi Yi Security

When Shuffling Fails: Token Recovery from Obfuscated KV Caches

LLM serving systems store, offload, and share KV caches, exposing token-level representations to infrastructure operators and co-tenants. Shuffled obfuscation combines secret linear transforms with fresh permutations, claiming security from b! possible shufflings. We show that input symmetry invalidates this counting argument: security depends on distinguishable outputs under chosen inputs. Repeated tokens produce identical first-layer Value rows, making every row…

Read at arXiv (research papers)

arXiv (research papers) Yavor Litchev Security

Retained Encrypted Search with Certified Split Joins

Retaining encrypted search across data edits and worker replacement requires separating computational validity from current release authority. We give exact Hamming search with at most one authorized decryption and decoding attempt per query preparation. Under authenticated prescribed execution, exclusive nonrollback owner state, an ideal public seed-expansion interface and a stated observation model, we reduce bounded adaptive privacy to ordinary ring learning with errors…

Read at arXiv (research papers)

arXiv (research papers) Rongxi Wang Security

PMKPIR: Parallel Multi-Keyword PIR with Interleaved Batching for Private Boolean Retrieval

Keyword Private Information Retrieval (Keyword PIR) allows users to retrieve data associated with specific keywords from a database while protecting query privacy. However, existing Keyword PIR schemes supporting boolean retrieval either incur high computational overhead due to multiple database traversals, or fail to support universal boolean computation. We propose PMKPIR, a novel keyword PIR scheme supporting both parallel multi-keyword queries and universal boolean…

Read at arXiv (research papers)

arXiv (research papers) Haoyang Liu Security

What Survives the Shift from Text to Pixels? Conditional Protection against Prompt Injection in Image-Rendered Documents

Multimodal language models increasingly answer questions over external documents, exposing them to indirect prompt injection: commands planted in untrusted content to hijack the response. Rendering untrusted content as page images lowers attack success on average, leaving open which injected operations survive rendering, at what utility cost, and whether failed image attacks reflect unread commands or commands read but not followed. We pair text and page-image versions of…

Read at arXiv (research papers)

arXiv (research papers) Jun Li Security

Erased but Recoverable: On the Fragility of Concept Erasure Under Fine-Tuning

Existing concept erasure methods for text-to-image diffusion models primarily focus on suppressing target concepts while preserving non-target generation quality. However, in practical applications, open-source diffusion models are often further adapted through DreamBooth, LoRA, or Custom Diffusion. Such adaptations can reshape internal representations, thereby weakening or even reversing the erasure effects achieved by existing methods. We identify erasure persistence as a…

Read at arXiv (research papers)

arXiv (research papers) Bonan Ruan Security

Guarding the Kingdom of Skill-Based Agents in the AI Dynasty of Computation: Control-Flow Integrity for Agent Skills

Control-flow integrity (CFI) confines a program to a control-flow graph fixed ahead of time. Procedural agent skills move this control flow into prompt text, and a large language model (LLM) agent decides at runtime whether each step runs in its reasoning, a tool call, or a script. The procedure's control skeleton survives, but its enforceability depends on where each step runs. Many violations, such as overwriting a file before validation, use only operations the skill…

Read at arXiv (research papers)

arXiv (research papers) Niccolò Lentini Security

PULSAR-V: I/O-Aware Post-Update Attestation for RISC-V

Post-update attestation proves that a device installed the intended firmware, not that the image stayed protected until the report was signed. In a multi-master SoC, a master outside the attested processor can corrupt update-critical memory. In contrast, the firmware identity, lifecycle state, and the CPU's protection evidence remain byte-identical to a healthy device's. We present PULSAR-V, a post-update attestation architecture which binds the authenticated update commit…

Read at arXiv (research papers)

arXiv (research papers) Long Xu Security

Evaluating LIM Trading Agents: A Critical Review of Backtesting Validity and Reporting Practices

Large language models (LLMs) are increasingly used in investment research, and many studies report promising historical returns. However, such results are only credible if researchers can demonstrate that the model used only information available at the time, account for how strategies and configurations were selected, and explain how reported returns were generated. We identify six common threats to validity in LLM investment evaluations and apply them to 18 published…

Read at arXiv (research papers)

arXiv (research papers) Wen Zhang Security

PIVOT: Distribution-Preserving Multi-BitWatermarking with Minimax-Optimal Scoring

As LLM-generated content becomes increasingly widespread, reliable attribution calls for multi-bit watermarks that embed recoverable messages without distorting the output distribution. We introduce PIVOT, a multi-bit watermark with model-free verification based on mirrored rank-interval sampling. At each generation step, PIVOT uses a keyed pseudorandom scalar and a secret permutation of the vocabulary. The message bit determines whether next-token probability intervals are…

Read at arXiv (research papers)

arXiv (research papers) Ziyuan Wang Security

Intent Contracts: Taking Tools Away from Agents

Intent contracts are an authorization interface for tool-using LLM agents in which the planning model holds no tools. For each step it submits a short logged contract: a goal, a verbatim quotation of the user, and scope fields. Code validates the contract, and a small trusted model checks it against the user's words and writes the tool call itself. Provided the validation and dispatch code is correct, every executed call is written by the trusted model, the planner's…

Read at arXiv (research papers)

arXiv (research papers) Alperen Örsdemir Security

When Do Decision Models Help LLM-Driven Network Discovery?

Network discovery, the identification of live hosts and their exposed services, is the first phase of every penetration test, and it is increasingly automated by harnesses driven by large language models (LLMs). To mitigate hallucinated or unsafe commands, recent work proposes inserting fast, non-generative decision models to validate actions prior to execution. This paper quantifies whether such validation improves discovery across three network scales (n = 10, 100, and…

Read at arXiv (research papers)

Older stories