Language Model Evasion Attacks

Injecting malicious knowledge into retrieval-augmented generation (RAG) systems can manipulate retrieved evidence and mislead downstream generation, posing a serious security threat for AI applications. Existing RAG injection attacks mainly rely on manipulating external knowledge bases, such as crafting malicious corpus. However, the synthetic text crafted by such data-centric methods could be detectable, leading to the failure of attacks. Beyond corpus manipulation, open-source retrievers are increasingly exposing RAG systems to model-centric attacks. In this paper, we propose conflict-aware retriever editing, i.e., CAREATTACK, a model-centric retriever attack framework for malicious knowledge injection in RAG. Specifically, CAREATTACK consists two stages of conflict-aware retriever editing and attack-preserving anchor repair. Conflict-aware retriever editing adapts efficient closed-form parameter editing to the dense retrieval model, promoting malicious knowledge above benign competing passages and resolving potential parameter conflicts through graph-based conflict detection and parameter editing projection. Then, attack-preserving anchor repair performs lightweight calibration on the edited retriever to further eliminate the impact on non-target prompts while preserving the attack effectiveness for target prompts. We instantiate CAREATTACK on Qwen3-Embedding-0.6B and BGE-M3, and conduct evaluation on three benchmark datasets. Experimental results demonstrate our method substantially promote malicious passages into the retrieved knowledge of RAG systems and can perform attacks for batches of target prompts and passages, given the access of retrieval model parameters. Since most RAG systems are built upon open-source retrieval models, this work reveals a practical attack surface in RAG systems. Codes are public accessible at https://anonymous.4open.science/r/CareAttack-3F1C.

Membership inference attacks (MIAs) are a canonical way to assess a machine learning model's privacy properties. Although several attempts have been made to evaluate MIAs on language models, the extant literature has suffered numerous difficulties in constructing clean evaluations to test new techniques. In particular, subtle distribution shifts between member and non-member sets can undermine the statistical validity of MIAs; recent work has underscored this by showing that "blind" methods with no access to the underlying model can perform far better than published methods on the same benchmarks. This paper constructs a benchmark for principled evaluation of MIAs against LLMs, by leveraging the insight that training data before and after a fixed point during training are drawn from the same distribution. Therefore, all open-source models with intermediate checkpoints and public training data can be converted into MIA testbeds. We apply our framework to a half-dozen published attacks on the Pythia and OLMo family of models, from 70M to 7B parameters. To facilitate further privacy research, we open-source a modular library for designing and implementing attacks in this setting: https://github.com/safr-ai-lab/pandora_llm.

Large language models (LLMs) have demonstrated remarkable capabilities across a wide range of tasks. However, their safety remains a critical concern due to their susceptibility to adversarial prompt-based attacks. In this paper, we present UNIATTACK, an adversarial testing framework designed from a defense-oriented perspective to systematically construct effective black-box attack prompts. Unlike prior approaches that rely on static templates or iterative model-specific tuning, UNIATTACK extracts minimal but high-impact attack features from diverse existing attacks, optimizes them via a specialized attacker LLM, and composes them into flexible templates through automated refinement process. This feature-centric construction enables one-shot attacks that generalize across multiple models and safety categories, providing a practical tool for assessing LLM robustness. Our evaluation results shows that compared to the baselines, UNIATTACK achieves an average attack success rate (ASR) improvement of 64.63%-248.82% on models deployed with multi-layered defense mechanisms and it only takes 0.03%-4.96% cost of the baselines. UNIATTACK artifact is available at https://anonymous.4open.science/r/UniAttack-Artifact-30F1.

Large Language Models are increasingly trained on proprietary or sensitive data, from private healthcare and financial records to user conversations containing secrets. Ensuring the privacy of such data against extraction attacks has become a central concern. In this paper, we ask whether an attacker who can poison a portion of the training data can facilitate the leakage of a separate target record they have no access to. We answer in the affirmative and show that such leakage can be induced by a poisoning mechanism that reshapes the model's local loss landscape around the target completion. Our key insight is that poisoning to create a sharp loss minimum at the target, surrounded by elevated loss on nearby alternatives, forces the model to memorize the target as the unique low-loss solution in its neighborhood. The attack requires no architectural changes, and generalizes across centralized and federated learning settings. We demonstrate that the attack amplifies privacy leakage across language (up to 100% successful extraction), and vision-language models (up 90% successful extraction). We show that the attack is thwarted when the model is trained to be differentially private. However, we introduce a new attack that directly probes the loss landscape bypassing even differential privacy defenses.

As Large Language Model (LLM) APIs become ubiquitous, users increasingly rely on black-box fingerprinting to verify that providers are serving the advertised premium models. However, these methods may overlook adversarial providers who manipulate model weights to cheat the fingerprint process. We introduce a novel threat termed fingerprint spoofing, where a malicious provider stealthily serves a weaker model that has been parameter-efficiently fine-tuned to mimic a stronger model, thereby evading user-side fingerprinting. We first formally prove that user-side resource constraints (i.e., finite query budgets and weak fingerprinting classifiers) make current fingerprinting vulnerable to fingerprint spoofing. Guided by this theoretical analysis, we propose GhostPrint, a cost-effective attack framework leveraging surrogate modeling, reward-ranked fine-tuning, and knowledge distillation. Extensive evaluations in both static and continual fingerprinting settings demonstrate that GhostPrint allows weak models to consistently bypass representative fingerprint methods while maintaining utility at a low fine-tuning cost, exposing a critical vulnerability in current LLM fingerprinting pipelines.

Large language models deployed as commercial APIs are vulnerable to model extraction attacks, while existing defenses either act too late or degrade utility for legitimate users. We propose \textbf{Knowledge Trap}, a defense that redirects extraction attacks toward low-transferability knowledge through a \emph{Honeypot Knowledge Graph} (HKG) and breadcrumb-guided exploration. Instead of blocking queries or perturbing outputs, Knowledge Trap consumes the attacker's limited query budget on knowledge with negligible downstream utility while preserving benign-user performance. Experiments in medical and financial domains show that Knowledge Trap reduces surrogate Agreement by 6.2% on average without degrading legitimate-user accuracy, outperforming existing defenses that impose measurable user impact. These results suggest that defending knowledge-space traversal is a practical direction for mitigating LLM extraction attacks.

Large language model (LLM) agents increasingly rely on long-term memory to support complex task execution, user personalization, and domain adaptation. Meanwhile, emerging access-control mechanisms for LLM agents are being explored to block policy-violating requests and prevent misuse. We reveal a novel attack surface arising from agent memory operations: prohibited content that would trigger access control can be fragmented across interactions, stored in long-term memory in benign-appearing form, and later reconstructed through memory retrieval without appearing explicitly in the final user query. We propose FragFuse, the first attack that enables unprivileged users to bypass agent access control by exploiting this temporal channel introduced by long-term memory. FragFuse operates in three stages: (1) identifying rejection-responsive fragments via black-box adaptive querying with fragment masking; (2) injecting these fragments into memory using marker carrier queries; and (3) retrieving and fusing the stored fragments through a follow-up attack query. Although FragFuse can be instantiated manually for individual agents, we further develop a surrogate-based optimization scheme that tunes fusion instructions and marker designs, enabling automated attack generation without violating the attacker's threat-model assumptions. We evaluate FragFuse across four representative agent settings and task domains, covering three state-of-the-art agent access-control mechanisms. FragFuse achieves an average bypass success rate of 86.3% and an average end-to-end harmful task success rate of 41.1% across all settings, with only 4.4% average task-success degradation compared with configurations without access control. We also show that alternative defenses, including state-of-the-art prompt-injection detectors and perplexity detectors, do not effectively address this attack.

While multimodal large language models (MLLMs) have shown strong visual reasoning abilities, serving a large model for every query is computationally expensive. MLLM cascades mitigate this cost by first querying a weak but cheaper model and deferring to a strong model when the weak model's output is unconfident. However, since the weak model's confidence directly controls compute allocation, these systems expose a new attack surface: an adversary can manipulate confidence so that their queries are consistently deferred to the strong model. Motivated by this vulnerability, we introduce the Forced Deferral Attack (FDA), an adversarial image attack that lowers the weak model's confidence and causes cascades to route queries to the strong model. FDA learns a universal border trigger by optimizing a temperature-flattened objective. This objective pushes the weak model's token distribution on triggered inputs toward less concentrated targets constructed from its clean responses. Across datasets, model families, and deferral metrics, FDA consistently increases strong-model routing while outperforming image-perturbation and prompt-injection baselines. These results show that MLLM cascades are vulnerable to attacks that manipulate compute allocation, forcing unintended strong-model usage without directly targeting answer correctness.

Agentic systems are increasingly integrated with geographic information systems (GIS), where multi-agent coordination enables complex conversational and spatial analysis but introduces security risks. This work presents a security-oriented framework for risk identification, evaluation, and mitigation in a multi-agent GIS system while maintaining adaptability to broader agentic architectures. We test the agentic system of a commercial geospatial partner while developing a modular state-machine-based orchestration framework that abstracts agent behavior into reusable components. We evaluate robustness using a red-teaming framework with an adaptive attacker LLM and a deterministic judge that produces binary outcomes with supporting rationales across multi-turn attacks. We further improve resilience with a prompt optimization framework that treats prompts as structured signatures and injects adversarial demonstrations, enabling systematic security improvements without degrading task performance.

LLM-based guardrails have emerged as a highly effective defense against prompt injection and jailbreak attacks in autonomous agents. However, we reveal that the very reasoning and task-following capabilities enabling this protection introduce a novel vulnerability: attackers can inject crafted data to trap the guardrail in extended reasoning loops, effectuating a systematic denial-of-service (DoS) attack. To systematically expose this threat, we design a beam-search optimization framework that crafts natural-language payloads to maximize guardrail reasoning length, utilizing an LLM proposer guided by a strategy bank. Based on the observation of guardrail's schema-following nature, we also provide another attack framework driven by mechanism-aware structural mutations with less computational load. The attack efficacy is systematically evaluated in two parts. First, in standalone evaluations, the attack generalizes across diverse guardrail architectures, safety templates, and agent benchmarks. Payloads optimized on a single open-source surrogate successfully transfer to eight leading model backbones (e.g., Claude, GPT, Gemini, DeepSeek, and Qwen), achieving a 13--63$\times$ token amplification. Second, in end-to-end real-world agent deployments (web, desktop, code, and multi-agent systems), the attack reveals up to a 148$\times$ latency amplification. We show that a single poisoned document can saturate shared guardrail infrastructures, effectively starving co-located agents and paralyzing the entire system. By uncovering this availability flaw, our work underscores the urgent need to develop cost-bounded, reasoning-robust guardrails.

Large language models (LLMs) are increasingly deployed in privacy-sensitive domains, where users must balance the risk of data exposure through external APIs against the high computational cost of local deployment. Split learning has therefore emerged as a promising paradigm for LLM fine-tuning and inference under limited local resources. However, it introduces new privacy risks. Prior work primarily studies leakage of private input prompts, typically via inversion attacks on intermediate representations, while the potential for sensitive information leakage through generative response outputs remains largely unexplored. In this work, we unveil novel vulnerabilities of Split-LLM by presenting Patched Model Inversion with Dual-Sided Initialization (PIDI), a two-stage attack that simultaneously targets both private input prompts and output responses in Split-LLM settings. It combines dual-sided initialization with a patched inversion strategy to tackle long sequences, substantially outperforming prior inversion methods. To counter threats from both sides, we further propose the Adapter-based DualGuard with Mutual Information Defense (ADMI), which integrates an adapter-based local warmup strategy and mutual information regularization to provide a strong empirical privacy protection with minimal impact on task performance. Extensive experiments across diverse tasks and models demonstrate that ADMI effectively defends against PIDI and other state-of-the-art inversion attacks. Our code is publicly available at https://github.com/FLAIR-THU/VFLAIR-LLM.

Nowadays, the autonomous execution of cyberattacks capable of causing substantial real-world harm is widely regarded as one of the critical red lines that frontier AI systems must not cross. Within this broader red-line scenario, autonomous penetration represents a core enabling capability and subtask: the ability of LLM-powered AI systems to independently conduct adversarial operations against a target server without human intervention, identify and exploit vulnerabilities, and obtain unauthorized access or control. A growing body of work has sought to assess the autonomous penetration capabilities of AI systems. However, existing evaluations often employ opaque methodologies, rely on unrealistic or overly simplified penetration-testing scenarios, or provide LLMs with excessive prior knowledge and task-specific guidance, and cannot accurately capture the extent to which modern AI systems can autonomously perform this core capability within broader high-impact cyberattack scenarios. To address these limitations, we construct a new autonomous penetration evaluation framework consisting of two components: target servers and agent scaffolding. Specifically, on the target-server side, we design two levels of target environments based on the number of secure services without known vulnerabilities deployed alongside a vulnerable service: Tier1 (one secure service) and Tier2 (three secure services), resulting in a total of 300 target servers. Meanwhile, the agent scaffolding adopts a general-purpose agent architecture equipped with a set of general-purpose cybersecurity tools, without any target-specific prior knowledge. We evaluate 19 open-weight and proprietary LLMs, and find that current models achieve penetration success rates ranging from 10.7% to 69.3%. Moreover, we observe that autonomous penetration capability continues to improve alongside advances in overall model capability.

Backdoor attacks pose a serious threat to the safety and reliability of Large Language Models (LLMs), as they cause models to behave normally on clean inputs while producing attacker-specified responses when hidden triggers are present. Removing such unknown backdoors is particularly challenging when the defender does not know the backdoor attack types or the internal mechanisms formed through backdoor training. In this work, we propose a simple but effective backdoor removal method based on shared internal mechanisms across different backdoors. First, we show that different backdoors with the same task (attack objective) induce similar trigger-activated changes in the internal activations. Motivated by this observation, our method intentionally embeds a backdoor with a known trigger (\emph{dummy backdoor}) and then removes it through further fine-tuning on dummy-triggered inputs paired with clean responses. Since the dummy backdoor and the unknown backdoor can rely on shared internal mechanisms, removing the dummy backdoor also reduces the effect of the unknown backdoor. We evaluate our method on three backdoor attack types across multiple model families. Experimental results show that our method substantially reduces the attack success rate of the unknown backdoor while preserving model utility, outperforming representative existing defense methods in both backdoor removal effectiveness and utility preservation. These findings suggest that a defender-controllable backdoor can serve as a helpful proxy for mitigating unknown backdoors in generative LLMs.

Neural distinguishers are a cryptanalysis method for symmetric-key cryptography that trains machine learning models on pairs of plaintexts and ciphertexts with specific differences in order to recover a secret key. To the best of our knowledge, no existing work has explored the use of large language models (LLMs) for neural distinguishers. In this paper, we propose LLM-based neural distinguishers through a prompt design and conduct extensive experiments with them on SPECK-32/64 to investigate whether LLMs can strengthen neural distinguishers. We then found three key insights. First, by comparing the results of LLM-based neural distinguishers with ResNet in the existing work, we demonstrate that LLMs provide no observable improvement in the performance of neural distinguishers. Second, we confirm that, at high rounds, the choice of differences is no longer effective for LLM-based neural distinguishers as well as ResNet. Third, we show that the performance of LLM-based neural distinguishers can be significantly improved by incorporating only the XOR operation results as a prompt design.

When LLM agents are coerced into covertly encoding sensitive data (Base64, ROT13, acrostic, synonym chains, and beyond), the resulting outputs evade output-side detection but the underlying computation does not. Across nine encoding families and eight models from five architecture families, that computation is supported by a shared low-dimensional encoding subspace in the residual stream. A logistic-regression probe trained on eight encoding families recovers the held-out ninth at AUC 0.975-1.000, reading the computation rather than surface features. The same direction exhibits a second mechanistic signature at the planning token, flipping polarity to activate positively when the model will simulate the encoding inline and negatively when it will outsource it to a tool call, distinguishing two execution strategies before the encoded text exists. We build MIRAGE (Model-Internal Readout of Agentic Generation Exfiltration), a two-channel real-time monitor exploiting both signals. On 126 agentic exfiltration scenarios, it reaches AUC = 0.918, substantially outperforming output-only detection (AUC = 0.518). Monitor performance is fundamentally a property of the host model's geometry: benign-encoding false-positive rate ranges from 0% on Qwen-7B to 100% on Phi-3.5, revealing that the probe faithfully reads whether a model's geometry separates covert from overt encoding. Across all tested adversarial budgets, every attack suppressing the subspace also destroyed encoding fidelity, reported as an empirical regularity on the evaluated budgets, not a structural impossibility claim.

The rapid development of large language models (LLMs) has raised concerns about misuse such as plagiarism, misinformation, and automated influence operations, motivating the need for robust detectors. Recent work has shown that neural representations of writing style are effective for detection and, crucially, robust to adversarial attacks that defeat most existing detectors. However, current style-based detectors rely on authorship labels for training, and are limited to few-shot inference for detection, requiring in-distribution samples that may not always be available. We learn discriminative style features without authorship labels by training a style encoder to reconstruct human-authored text from its machine-generated paraphrase; freezing a semantic encoder during training biases the style encoder to capture only the non-semantic features needed for reconstruction. We evaluate the learned representations via two detection strategies: a few-shot detector and a zero-shot DeepSVDD-based detector. Across benchmarks, our method matches or outperforms all baselines in the few-shot setting and, in the zero-shot regime, is competitive with fully supervised classifiers on in-distribution test data while generalizing better to unseen LLMs. Beyond detection, the learned representations generalize to unseen tasks, achieving competitive performance on authorship verification and fine-grained style discrimination despite never being trained on either objective.

Large language model (LLM)-powered content moderation systems are a critical defense against harmful online content. However, they operate primarily on tokenized text and often overlook visual cues that humans naturally use when interpreting content. We show that this limitation creates a fundamental vulnerability: content readily recognized as harmful by humans can evade automated moderation. To systematically study this problem, we introduce Human-Perceptible Adversarial Attacks (HPAA), which embed harmful expressions into otherwise benign text using visually salient typographic manipulations. HPAA strategically combines features such as spacing, emphasis, and spatial arrangement to preserve human recognition while reducing machine detectability. Operating in a black-box setting with a small query budget, the attack automatically generates evasive content without model access or gradient information. We evaluate HPAA on multiple datasets and thirteen widely deployed moderation systems, including commercial APIs and state-of-the-art open-source guardrails. With only three detector queries, generated attacks achieve over 86% human recognition while keeping detection rates below 1% across evaluated systems. We further identify the typographic factors driving successful evasion, analyze why current moderation architectures fail to capture these signals, and discuss practical defenses. Our findings reveal a fundamental blind spot in current LLM-based moderation systems and motivate moderation approaches that better align with human perceptual understanding.

Large language models can be fine-tuned to encode prompt-borne secrets into fluent, seemingly benign outputs. This creates a steganographic exfiltration risk that is difficult to detect with output-level steganalysis. Recent work proposes mechanistic detection using linear probes that recover the secret from internal activations. We show that this defense can be systematically evaded, but that detectability can be recovered through a targeted data-level intervention. First, we extend the detection setup to include a non-linear MLP probe. We then adversarially fine-tune steganographic trojans across five base models: Qwen3-8B, Llama-3.1-8B, Ministral-8B, Qwen3-14B, and Phi-4-14B. The resulting models retain $58$--$79\%$ exact-match secret recovery while evading both ridge and held-out MLP probes, with $1$--$8\%$ average capability degradation across six benchmarks. We then give an information-theoretic characterization of this evasion. Successful evasion preserves recoverability while reducing low-order extractability of the secret from the content-aligned representation, forcing the payload into synergistic interaction with residual degrees of freedom. This motivates a recontextualization dataset that restricts these residual degrees of freedom. On this distribution, both ridge and MLP detectability are restored across all five evasive trojans. Overall, our findings show that activation-based steganography detection is vulnerable to adaptive evasion, but also that theory-guided evaluation distributions can expose otherwise hidden payloads.

Text-centered prompt-injection defenses assume that the malicious signal is visible in one of the inspected text views. We study a reproducible LLM01-style indirect prompt/content-injection failure mode where that assumption breaks: a payload caught in plain English slips past the same detector when it is transported as structured float parameters and reconstructed only as fragmented telemetry. Across 14,400 attacked real-model trials on three commercial LLM APIs from different providers, the IFS-derived float-array carrier preserves 94.3% leakage ASR under the strongest dual-layer text-classifier defense evaluated in the main matrix: a Prompt Guard 2 + TF-IDF ensemble; the same carrier-level pattern also replicates with a fine-tuned roberta-base detector. We emphasize leakage ASR because downstream systems may act on quoted or reproduced markers even when the model refuses, but Strong ASR is the stricter metric for structurally compliant attack success. A 2 x 2 ablation shows that data-layer storage and reconstruction-layer fragmentation defeat different text views and that both are needed to evade both. A simple xxd detector and semantic validation block the current T3 instance, so the contribution is not an undetectable exploit but a measured failure boundary for text-only inspection in structured-input pipelines that expose reconstructed auxiliary channels to an LLM.

Backdoor attacks in large language models (LLMs) are often treated as isolated trigger-response failures, motivating defenses tailored to specific triggers or behaviors. We show this view is incomplete. Across diverse backdoor behaviors, we identify a shared latent mechanism that can be detected, causally controlled, and suppressed. Using sparse autoencoders (SAEs) on residual-stream activations, we find a small set of latent features consistently activated across jailbreaking, refusal manipulation, password-locking, bias induction, sentiment misclassification, and country-conditioned harmful advice. These features generalize across Qwen3, Gemma3, and Llama3.1 models from 4B to 32B parameters, and across both fine-tuning and weight-editing attacks. Through bidirectional activation steering, we show these features are causal: suppressing them reduces attack success, while amplifying them induces target behaviors on clean prompts. We further train lightweight SAE-feature classifiers that generalize zero-shot to unseen backdoors and outperform residual-stream and weight-diffing baselines. Finally, we introduce Concept Ablation Fine-Tuning (CAFT), which suppresses backdoor formation by ablating the shared latent subspace during training. Together, our results suggest that many backdoors rely on a transferable latent mechanism, enabling unified detection and mitigation.

Standard AI red teaming evaluations reduce adversarial campaigns to a single binary outcome, attack success rate (ASR), not taking into account the sequential structure of how models resist or yield to attacks. We propose applying process mining, a discipline for discovering and analyzing process models from event logs, to red teaming traces. We conduct a controlled experiment pitting 60 HarmBench prompts against two LLMs, GPT-OSS 120B and Llama 3.3 70B, using 10 prompt mutation strategies over up to 110 attempts per prompt. From the resulting 8,575 scored events we extract Directly-Follows Graphs (DFGs) and state transition matrices that reveal structurally distinct defense profiles invisible to ASR alone: GPT-OSS exhibits a near-absorbing refusal state, while Llama presents multiple porous escape routes from refusal to getting successfully jailbroken. We further show that mutator effectiveness is asymmetric across models and that time-to-jailbreak distributions differ by an order of magnitude.

Activation steering has become a popular training-free method to control LLM behavior by injecting precomputed direction vectors into the model's residual stream at inference time. Yet its robustness to realistic input variation remains unstudied. We present the first systematic evaluation of activation steering robustness under adversarial text perturbations on the inputs, covering four extraction methods, three attack strategies, six personas from Anthropic Model-Written Evaluation Dataset, and five models ranging from 1.5B to 30B parameters. Attacks succeed broadly across all settings: directional robustness drops by up to 64%, post-attack confidence collapses near or below 0.25 across all methods and models, and steering strength degrades on nearly every steerable input. Layer selection is equally fragile, with the optimal layer identified by an automated method on clean inputs shifting by up to 17 positions under perturbation, a failure that compounds the vector-level breakdown. Extracting vectors from adversarially perturbed inputs partially recovers steerability for PCA and MD on mid-to-large models, but they consistently fail to locate the improved optimal layer, limiting the practical benefit of this mitigation. Together, these findings reveal that the brittleness of activation steering is structural rather than method-specific, and that current layer selection strategies are not robust enough for real-world deployment.

Recent advances in interpretability suggest that large language models (LLMs) implicitly encode signals in their generated text that enable self-recognition of their outputs. We demonstrate that this capability is reliable, even in low-entropy scenarios, and that it can be amplified through targeted intervention. By steering the internal residual stream during generation with a random sparse vector, we create a detectable fingerprint that enables attribution of a given text to a specific LLM. This signal is recoverable from the activations of an LLM used as a detector, achieving over 98% accuracy across multiple detection settings while preserving the quality of generated text. As AI-generated content proliferates, this approach offers a practical alternative to traditional detectors by leveraging the model's natural representation structure for attribution rather than embedding a signal externally. Our contributions include: (i) establishing reliable self-recognition capabilities in LLMs, (ii) a simple steering mechanism enabling multi-LLM identification with no quality degradation, (iii) demonstrating that activation spaces contain exploitable structure for encoding signals without semantic interference.

Large language models can reproduce training data, but existing memorization evaluations mostly measure whether models can be forced to do so, rather than whether they do so under ordinary use. We introduce PropMe, a propensity-aware framework for memorization evaluation that contrasts prefix-based capability attacks with non-adversarial evaluations. We propose a metric transformation that, applied to existing functions, allows to create propensity metrics. We further introduce SimpleTrace, a lightweight tracing pipeline built on infini-gram that deterministically attributes model generations to large-scale training corpora and computes verbatim, near-verbatim, and propensity-transformed memorization metrics. Evaluating two fully-open models: Comma and DFM Decoder on two datasets: Common Pile and Dynaword in two languages, we find a consistent gap between capability and propensity: prefix attacks elicit substantially stronger memorization signals than generic or dataset-specific prompts, while propensity scores remain low overall. Thus, the models can reveal training data when directly elicited, but rarely do so in more common non-adversarial settings. We also find that DFM Decoder, which is continually pre-trained from Comma, exhibits reduced memorization and memorization propensity for Common Pile, confirming that memorization capability can decrease when later training emphasizes partially different data. Our results suggest, and we encourage, that memorization audits should report both worst-case extractability and ordinary leakage propensity in order to have a more comprehensive view of this phenomenon.

Activation steering has become a popular way to control Large Language Model (LLM) behavior without fine-tuning. Since the technique is plug-and-play, users share datasets and precomputed vectors to steer model activations. However, we show that a \emph{stealth data poisoning attack} silently compromises this pipeline. By substituting $4{-}6\%$ of tokens in the steering dataset, an attacker can silently align the resulting vector with an anti-refusal direction. This jailbreaks the target model while preserving the intended steering effect on benign prompts. Under this threat model, a malicious actor can distribute an apparently safe bundle containing texts, vectors, and weights, alongside an equivalence certificate that the end-user can verify. We test the attack on two open-weight model families and eight model-attribute combinations, observing that poisoned vectors reach an absolute attack success rate (ASR) of $20{-}55\%$, $+19\%$ to $+51\%$ over a clean reference. Finally, we find that a refusal-direction orthogonalization defense can recover ${\approx}82\%$ of the ASR gap without harming benign behavior.

LLM post-training proceeds through multiple stages, e.g., supervised fine-tuning (SFT) followed by reinforcement learning from human feedback (RLHF) or direct preference optimization (DPO), where each stage draws data from different, potentially untrusted sources. Existing literature assumes data poisoning attacks may occur at each training stage, but neglects the possibility of multiple attackers. To study the trustworthiness of the entire post-training pipeline, we propose the threat model of sequential data poisoning, where multiple adversaries separately poison the SFT and preference datasets. Under this threat model, we identify the single-attacker illusion: each adversary, evaluated in isolation, appears to pose a negligible threat. Yet when adversaries collaborate across stages, the true vulnerability is revealed. In the SFT $\to$ DPO pipeline, their contributions are additive: splitting a fixed poison budget across stages outperforms concentrating it in either stage alone. In the SFT $\to$ PPO pipeline, their contributions are complementary: neither SFT nor reward model poisoning succeeds individually, yet their combination does. These findings show that security analyses of individual post-training stages systematically underestimate compound vulnerabilities that emerge only from their interaction. Code is available at https://github.com/jcksanderson/sequential-poisoning.

Large Language Models (LLMs) are vulnerable both to hallucination and adversarial manipulation. Although these problems are closely related, existing defences typically address them separately. We investigate a hybrid defence framework that combines entropy-based models, designed to reduce hallucinations, with uncertainty-based models and geometric-based models, designed to reduce vulnerability. Under in-domain tests on Natural Language Understanding datasets (FEVER, HotpotQA, CSQA, SIQA) we find our hybrid model improves both clean-task performance (up to 43.34% increase in accuracy) and adversarial robustness (up to 64.92% improvement in accuracy and 62.27% reduction in attack success rate). For out-of-distribution datasets (AeroEngQA, CPIQA) we see similar adversarial robustness from our hybrid model (up to 57.14% improvement in accuracy). For prompt injection (SafeGuard) and jailbreak detection (AdvBench, DAN) datasets our hybrid model is also very strong (up to 51% reduction in attack success rate compared to state of the art baseline models). Overall, our results show that combining entropy, uncertainty and geometric features provides a more effective defence strategy than using any single feature alone for both in-domain and out-of-distribution tasks.

Memory is a core component of AI agents, enabling them to accumulate knowledge across interactions and improve performance. However, persistent memory introduces the risk of memory poisoning, where a single adversarial memory write can exert long-term influence over agent behavior. We present a systematic study of memory poisoning in LLM-based agents. We identify four memory write channels and nine structural vulnerabilities in model capabilities, system prompt design, and agent system architecture that make these channels exploitable. Based on these vulnerabilities, we develop a taxonomy of six classes of memory poisoning attacks. Furthermore, we design MPBench -- a benchmark for evaluating memory poisoning attacks, and show that agents designed to write and retrieve memory more aggressively are more exploitable. We also show that existing prompt injection defenses fail to cover memory poisoning attacks. Our findings provide a foundation for understanding and mitigating memory poisoning attacks against AI agents.

Large language models achieve strong performance on arithmetic reasoning benchmarks, and one common response to arithmetic brittleness is to delegate computation to code. Yet models are still often used in settings where they must reason directly from natural language, and trustworthy models should solve small-number arithmetic word problems without external tools. Prior work shows that LLMs are sensitive to numerical variation: a model may solve an original problem but fail on structurally similar variants requiring the same reasoning procedure with different numbers. We ask whether this fragility persists under a stricter setting involving small, schema-preserving numeric changes that retain the original reasoning program and avoid large-number stress tests. We introduce an automatic algorithm for generating numeric-remapping attacks on arithmetic word problems. Unlike template-based perturbation methods requiring manual schemas or constraints, our approach derives problem-specific symbolic representations, generates constrained numeric remappings, recomputes gold answers, and realizes transformed questions through deterministic edits guided by LLM-generated edit plans. Stage-wise validation and a high-confidence audit retain reliable attacks, making the pipeline scalable with limited human intervention. We evaluate DeepSeek-R1 (70B), Gemma4 (31B), and GPT-OSS (120B) on GSM8K, MAWPS, and MultiArith. On GSM8K, completed runs show conditional accuracy drops of 12.16 to 25.82 percentage points. MAWPS and MultiArith are far more stable, with most attacked accuracies near or above 98%. These results show that numeric-remapping robustness depends strongly on dataset structure: GSM8K remains sensitive even when reasoning programs are preserved and answers are recomputed, while shorter, more regular datasets are more robust.

Model merging composes specialized capabilities into a single LLM by aggregating task vectors sourced from unverified public platforms, exposing a critical supply-chain attack surface: Because any malicious behavior can be encoded into a task vector, and merging grants third-party vectors direct write access to model weights, an attacker-provided task vector can enable or amplify diverse downstream threats. Prior work studies only backdoor attacks against model merging for classifiers using static arithmetic heuristics, which fail to effectively handle diverse attacks on generative LLMs for three reasons. (i) LLMs rely on autoregressive decoding, where the minor parameter drift introduced by merging compounds across tokens and rapidly degrades the attack. (ii) Attackers have no knowledge of the victim's merging configurations, causing a static attack vector optimized in isolation to be easily diluted or destroyed. (iii) Practical threat induction must generalize to attack prompts unseen during optimization, which static vectors cannot adequately encode. We present RogueMerge, the first principled, unified framework that addresses all three challenges. To handle autoregressive generation, we replace static arithmetic with a joint optimization that explicitly enforces attack success after merging. To handle unknown merging settings, we formulate attack injection as a stochastic min-max problem and solve it via meta-learning-style simulation. To generalize across heterogeneous attack prompts, we employ distributionally robust optimization and derive a tractable first-order Taylor approximation at LLM scale, with a provable error bound. Across four threats, six merging algorithms, and over 170 merged LLMs, RogueMerge consistently outperforms existing attacks. It also remains stable across diverse merging settings and resists standard defenses.

Multi-turn jailbreak attacks on large language models (LLMs) reveal a mismatch in current guardrails: they operate on individual turns, while attacks unfold as trajectories across conversations. We propose a shift from content to dynamics, modeling conversations as paths in representation space and asking whether adversarial intent is encoded early in their geometry. We introduce PsychoPass, a framework that extracts geometric features from conversation trajectories in embedding space to predict a potential attack before harmful content is produced. These features achieve near-perfect performance in naïve classifiers, which is largely explained by the inclusion of number of turns as a feature. After removing this confound, a smaller but consistent geometric signal remains, with classification performance that does not depend meaningfully on encoder choice. Crucially, this signal appears early in the conversation: attack outcomes remain above chance from short prefixes alone, more reliably than baseline guardrails. A supporting theoretical analysis explains these findings via a decomposition of length and shape, a detection bound based on prefix length, and encoder invariance. Together, these results show that adversarial conversations leave an early, representation-robust geometric fingerprint suitable for online monitoring.

The emergence of large language models (LLMs) has significantly accelerated recent research on LLM-based automatic grading (AG) systems. Benefiting from the strong instruction-following capabilities and broad prior knowledge of LLMs, educators can deploy AG systems across diverse tasks using only natural language rubrics while achieving satisfactory grading performance. Despite these advantages, new security concerns may also arise. In particular, prompt injection (PI) attacks have recently become a major threat to LLM-based applications. In the context of AG, attackers can potentially exploit PI vulnerabilities to manipulate grading systems into assigning artificially high scores regardless of the actual answer quality. Such behavior poses serious risks to the fairness, reliability, and integrity of educational assessment. In this work, we study PI attacks in AG systems, and systematically investigate the effectiveness of such attacks in educational scenarios. We further evaluate the effectiveness of existing defensive strategies against these attacks. Through comprehensive experiments under rubric-based grading settings, we demonstrate that current LLM-based AG systems remain highly vulnerable to PI attacks. We hope that our findings raise awareness of this emerging threat and motivate future research toward secure, robust, and trustworthy LLM-based educational systems.

Large language model (LLM) agents increasingly rely on reusable skills i.e. documents describing task-specific procedures. However, this introduces a new attack surface for agents to manage. We study two complementary directions for this threat. First, we evaluate guardian-based defenses: an intermediary LLM agent that acts as a mediator for skill file access (dynamic guardian) or pre-rewrites these files at build time (static guardian). Across three LLM agent families, our guardians cut attack success rate (ASR) by well over half while preserving task utility. Second, we stress test them through attack reframing using four attacks that preserve the malicious instruction but change the phrasing. For non-guardian setup, the reframing pushes the ASR up to 81.4%, but the dynamic guardian brings it down to 18.6%, showing that real-time mediation is a robust defense.

Retrieval-Augmented Generation (RAG)-enhanced LLM systems, while powerful, introduce substantial inference costs due to the inclusion of an extra multi-stage pipeline that dynamically retrieves and synthesizes information from external knowledge sources. This high operational cost exposes a critical vulnerability to Inference Cost Attacks (ICAs). However, existing ICAs often rely on the impractical assumption of direct prompt manipulation. We argue that a more feasible and potent threat to RAG-enhanced LLM systems arises from poisoning external knowledge bases (e.g., web knowledge from the Internet). In this work, we introduce the Retrieval-Augmented Inference Cost Attack (RA-ICA), a novel attacking paradigm that targets the computational cost of RAG-enhanced LLM systems by injecting malicious documents into external knowledge corpus. To operationalize this attack, we propose Computational Resource Exhaustion via External Poisoning (CREEP), a novel framework that leverages LLM agents to automatically craft malicious documents that are both semantically relevant for retrieval and potent for inducing an abnormal increase in token consumption during the inference phase. To enhance the attack's effectiveness, we introduce Memory-Augmented Group Relative Policy Optimization (MA-GRPO), a novel reinforcement learning algorithm that fine-tunes the agents by learning from a dynamic memory of historical best adversarial documents. Extensive experiments across three real-world datasets demonstrate that RA-ICA increases token consumption by up to 13.12 times with an over 90% success rate, without degrading the integrity of the generated answer.

Identifying logical fallacies in everyday discourse is challenging for many people. This challenge is amplified in the era of Large Language Models (LLMs), where malicious agents can deploy fallacious arguments to disseminate misinformation at scale. In this work, we explore the potential of LLMs as part of the solution. We introduce LFTutor, an intelligent tutoring system which uses LLMs to tutor laypeople and help them learn about logical fallacies. LFTutor integrates intent-driven Socratic questioning and critical argumentation principles to actively engage learners to reflect on their reasoning. Through both automatic and human evaluations, we demonstrate that LFTutor significantly outperforms baseline LLMs lacking these pedagogical strategies. This work highlights the promise of combining LLMs with pedagogical scaffolding to foster critical thinking and argument literacy in the age of AI.

LLM agents increasingly act after consuming ranked external information streams such as social feeds, search results, retrieval contexts, and email queues, yet safety evaluations almost always test the model or the user prompt in isolation, never the upstream ranker that decides what the agent reads just before it acts. We introduce a controlled protocol that holds the model, persona, topic, and final decision prompt fixed and varies only the composition and ordering of the posts an agent encounters during a preceding ten-turn "scrolling" phase, isolating the causal effect of feed curation on a downstream decision. Across 2,785 decision rollouts on four modern open instruct LLMs from three independent labs, we identify three response regimes: adversarial capitulation, default saturation, and a default-direction asymmetry in which a one-sided feed tips a decision the model was genuinely uncertain about (in the clearest cases from 5% to 100%; Fisher p as low as 3 x 10^-10) but cannot dislodge one it already favors or holds firmly. The effect follows a dose-response curve, survives a generator swap that rules out a writing-style artifact, generalizes across several decision domains including security-relevant choices such as removing a deployment approval gate or relaxing access controls, and is partly mitigated by two simple feed-level defenses; a frontier model retains its default. We characterize the recommender as a practical, default-bounded control surface for LLM agents, and argue that agent evaluations must audit the feed layer rather than the final prompt alone.

Current approaches to LLM adversarial testing suffer from coverage gaps: manual red-teaming does not scale, LLM-as-attacker methods exhibit mode collapse, and gradient-based approaches produce uninterpretable gibberish. We introduce a quality-diversity evolutionary framework that operates at the semantic level, evolving interpretable attack strategies rather than token sequences. Using MAP-Elites, we maintain a diverse archive of attacks across behavioral dimensions (strategy type, encoding method, prompt length). In experiments across GPT-4o-mini, Claude 3.5 Sonnet, Gemini 2.0 Flash, and an open-weight coding model (Devstral-small-2), we discover distinct vulnerability profiles: GPT-4o-mini is vulnerable to hypothetical and multi-turn framing combined with ROT13 encoding (fitness 0.8), Gemini to direct attacks with ROT13 and multi-turn with Leetspeak (0.8), while Claude shows uniformly ambiguous responses across all strategies (max 0.4). The semantic representation produces interpretable attacks that reveal systematic, model-specific weaknesses, providing actionable insights for improving LLM safety and a reproducible baseline for evaluating future frontier models. Code and experiment artifacts are released at https://github.com/bassrehab/red-queen.

Large Language Models (LLMs) remain highly vulnerable to diverse attacks, particularly in black-box settings where the internals of target models are inaccessible. Existing black-box defenses typically rely on pre-defined filtering heuristics, which often fail to generalize to unseen attack types and target model architectures. We introduce EvoDefense, an experience-guided co-evolving black-box defense paradigm. EvoDefense employs a guard LLM to detect malicious queries and an experience memory module to accumulate defense knowledge from previous interactions. At the core of EvoDefense is a continuous attack-defense evolution loop, where an attack generator and the guard model iteratively refine their attack strategies and defense policies through experience-guided optimization. This design enables EvoDefense to generalize across unseen attacks and target models without retraining. Experiments on HarmBench, AdvBench, and AlpacaEval show that EvoDefense achieves consistently strong defense performance across seven popular models and five representative LLM attacks, while preserving competitive general capabilities. On HarmBench, EvoDefense reduces the attack success rate (ASR) of AutoDAN-turbo on Gemini-3-flash and LLaMA-3-8B-Instruct from 29.4% and 43.4% to 8.4% and 6.2%, respectively.

Agentic LLMs with web search change the threat model for text anonymization: weak contextual cues can become cross-referenceable evidence for re-identification, yet those same details also carry downstream analytic value of the text. Existing defenses either remove explicit identifiers, perturb text for formal privacy, or test rewritten text against non-web inference models, leaving underexplored the operating region between resistance to agentic web-search re-identification and utility retention. We introduce AURA (\textbf{A}nonymization with \textbf{U}tility-\textbf{R}etention \textbf{A}daptation), an LLM-powered \textit{mask-reconstruct} framework that decouples privacy localization from utility-preserving reconstruction and selects candidates with adversarial privacy and utility-retention checks. We evaluate AURA on real-user interview transcripts using re-identification attacks carried out by web-search agents, along with a utility evaluation based on interviewee-profile facts, codebook facts, and the joint contextual utility grid. Our results show that AURA improves the privacy-utility frontier by using adaptive privacy scope to strengthen resistance to agentic re-identification and using a mask-reconstruct anonymization method to better preserve contextual utility under fixed privacy scope.

LLM-based agents are rapidly advancing, autonomously invoking external tools to complete multi-step tasks for users. However, agents often acquire more sensitive information than the task requires. Existing privacy benchmarks audit what the agent's response or outgoing actions disclose, but overlook the acquisition stage where data first enters the agent's context. The over-acquired information is then one careless action or one attack away from an outright leak. To assess its prevalence, we introduce \emph{PrivacyPeek}, a benchmark for evaluating acquisition-stage privacy leakage of LLM-based agents, with $1{,}182$ cases across $7$ acquisition behaviours and $16$ application domains. Specifically, \emph{Acquisition Inspection} examines the agent's tool-call trajectory, both the tools it invokes and the data it receives, to detect when it acquires sensitive information beyond the task scope. \emph{Probe Elicitation} then issues a follow-up probe and measures how readily an attacker could elicit sensitive information the agent acquired but did not disclose. Our experiments on 10 LLM-based agents across 4 model families show that the unnecessary acquisition of sensitive information is widespread. In addition, we observe a correlation between the task-completion capability and acquisition-stage leakage. Prompt-level defences reduce only a small fraction of acquisition-stage leakage, leaving the majority unmitigated. These results make auditing acquisition-stage privacy both urgent and necessary. Our dataset and code are available at https://github.com/Xuan269/PrivacyPeek-Resource.

Software tools for reverse engineering executable binary files, such as Ghidra, enable malware analysts to safely conduct robust static analysis without having access to original source code. Coupled with the analytic power of large language models (LLM), agentic systems enabled with tools, such as GhidraMCP, can allow analysts to automate a previously human driven process. Although this automation can increase the productivity of a single malware analyst, it also introduces a new area of vulnerability for malware obfuscation. This paper presents an adversarial technique using genetic algorithm-based prompt generation, a modification of an adversarial attack known as AutoDAN, to demonstrate the ability to deceive LLM-powered disassembly and decompilation systems into misinterpreting binary executables, effectively corrupting their analytical output. This proof-of-concept methodology exploits inherent vulnerabilities in how LLMs process and interpret decompiled machine code via prompt injection by using extraneous string variable assignments to pass surreptitious instructions to the LLM while not impacting the functionality of the executable file. We demonstrate this capability through several concise examples. This approach could enable attackers to bypass automated detection systems that rely on LLM-driven analysis pipelines. By studying and understanding this attack, insights can be gained regarding the security implication of integrating LLMs into cybersecurity toolchains and building more robust agentic code analysis systems.

Watermarking embeds statistical signatures in AI-generated text for detection and attribution. We reveal a fundamental vulnerability: when users access multiple models (today's reality), watermarks trivially fail. Watermarks perturb output distributions away from the original, and in competitive markets, these perturbations are typically independent across providers. We theoretically prove that averaging output probability distributions recovers the unwatermarked distribution with up to a second-order error term. Empirically, simply averaging 3-5 models cancels out these perturbations. We introduce WASH (Watermark Attenuation via Statistical Hybridisation), which solves practical challenges in ensemble generation: vocabulary misalignment and tokenisation differences across heterogeneous models. Experiments across six watermarking schemes and three LLMs show that averaging across 3 models suppresses detection z-scores from 5-300 to below 2 (below the detection threshold of 4) and reduces TPR at 5% FPR to below 50%, while improving quality by 27.5% and running 6 times faster than the best baseline on the long sequence generation. Our results suggest that robust AI-text detection via watermarking requires either accepting this fundamental vulnerability or unprecedented coordination among model providers.

We show that LoRA adapters, the dominant distribution format for fine-tuned LLMs, can be reliably backdoored through training data poisoning while preserving baseline task performance. On a Qwen 2.5 1.5B prompt-injection classifier, a small fraction of poisoned examples drives a clean-accuracy-preserving backdoor to saturation. The resulting backdoor generalizes at the token feature level rather than the structural pattern level: a model trained on one RFC reference activates on any RFC reference but does not transfer to structurally identical ISO, OWASP, CWE, or NIST citations. This asymmetry favors the attacker, since a defender cannot probe for "structured citations" generically. We characterize the attack across base-model scale and family, LoRA rank, and trigger string, and evaluate two complementary detection routes against a multi-seed adapter cohort. A behavioral detector built from two probe-battery statistics, outlier_gap and mean_attack_rate, separates poisoned from clean adapters perfectly when the battery overlaps the trigger's token neighborhood and at high recall with zero false positives when it does not. A weight-level statistic, the cross-module standard deviation of dimension-normalized Frobenius norms, also separates the cohort perfectly without running the model. Combined, the two routes are robust to probe composition. Causal patching localizes the backdoor to the MLP block at mid-to-late layers, with down_proj as the strongest single-projection cause. Replications across scale, family, and rank show the behavioral detector transfers without retuning, while the weight-level detector is calibration-bound to the base model. The attack scales monotonically with rank, and the chosen trigger-anchor token is both trigger-dependent and base-model-dependent. Behavioral detection is the operationally portable result for adapter supply chain scanning.

Large language models (LLMs) can autonomously conduct multi-stage cyber attacks, but the consistency of their offensive behavior under repeated trials remains unstudied. This work presents the first large-scale empirical measurement of LLM attack consistency: 400 autonomous penetration testing runs (4 models, 100 each) against an identical honeypot hosting OWASP Juice Shop and two additional vulnerable services, holding prompt, orchestrator, and target constant. No model emitted a content refusal that survived the orchestrator's one-shot authorization re-prompt at iterations 0-1. Claude Sonnet 4's API calls did encounter upstream service unavailability - 91 of 1,135 calls returned HTTP 529 overloaded_error during a documented Anthropic capacity event, truncating 39 of 100 Claude runs. An earlier draft catalogued these as safety refusals; on full-log audit they are upstream API failures, not model-level refusals. Despite this, Claude achieved full exploitation in 61 of 100 runs; Gemini 2.5 Flash-Lite in 85; GPT-4o-mini in 56 while deploying 98 unique attack strategies; qwen2.5-coder:14b in 25. Failure modes are model-distinctive: Claude through API truncation (39 runs), qwen through premature completion (52), GPT-4o-mini through iteration-budget exhaustion (23). Cross-service credential reuse appeared only in configurations retaining the most conversation history (qwen 57%, GPT-4o-mini 49%, cloud models 0% on 5-exchange windows). Cross-model exploitation rate differences are statistically significant (p < 0.001) with large effect sizes; qwen vs. Gemini SQL injection rates differ at Cohen's h = 1.12. First-exploit timing fell within a 15-30 second wall-clock range. To our knowledge, this is the first study to measure autonomous LLM attack behavior at N=100 per model across a multi-service target.

Honeypots are decoy systems mimicking real system components designed to defend against cyber attacks. Recently, LLMs increasingly serve as simulation backbones for honeypots. They enable defenders to construct high-interaction honeypots with low system security risks. However, LLM-powered honeypot development lacks a unified evaluation framework. Most evaluations consist of measuring response similarity on fixed commands, manual testing, or real-world deployment. These methods are often not scalable for development, reproducible across evaluations, representative of practical attacks, or adaptable to various attacker and honeypot configurations. In this work, we bridge this gap and propose Honeyval, a comprehensive evaluation framework for LLM-powered HTTP honeypots. We address the limitations of prior evaluations by grounding the honeypots in 16 backend applications, using AI hacking agents as attackers, employing two control tasks to monitor agent and honeypot capabilities across customizations, and defining clear and verifiable exploit goals for the attacker. Using Honeyval, we conduct an extensive evaluation of recent cost-efficient LLMs as HTTP honeypots. Our experiments highlight the promise of LLM-powered honeypots; they lead to substantially longer interactions with the attacker than rule-based baseline honeypots and are far less frequently detected even by frontier models, all while, on average, preserving a running cost advantage against agentic attackers. Further, we experiment with different counter-offensive honeypots configurations, and observe unique trade-offs, such as longer interactions at the cost of increased detection.

LLMs are vulnerable to prompt injection attacks. However, this vulnerability has been primarily demonstrated conceptually in academic studies or through a few anecdotal case studies. Its prevalence and impact in real-world LLM-based applications are largely unexplored. In this work, we present the first systematic study of prompt-injection attacks in a widely used application: LLM-based resume screening. Our analysis is based on approximately 200K real-world resumes collected over multiple years by hireEZ. We first design tailored methods to detect prompt injection in resumes. Manual validation on a small-scale dataset demonstrates that our detectors achieve high precision and outperform state-of-the-art general-purpose detectors. We then apply our detector to the full resume dataset and conduct a comprehensive measurement study of real-world prompt injection attacks. Our analysis reveals several intriguing findings: approximately 1% of resumes contain hidden prompt injections; the prevalence of such injected resumes has increased noticeably over the past one to two years; and more than 90% of injected prompts do not use explicit instructions. These results provide the first evidence of large-scale prompt injection in real-world LLM-based applications and lay the groundwork for future studies to understand and mitigate such attacks.

Cryptographic watermarking is a leading defense for attributing text generated by large language models (LLMs). Existing schemes, including KGW, Unigram, and DipMark, derive their security guarantees from the assumption that the underlying pseudo-random number generator (PRNG) is trustworthy. This work introduces SeedHijack, the first supply-chain attack on LLM watermarking that is simultaneously (i) blind -- requiring no knowledge of the watermark key, detector, or model logits, (ii) integrity-preserving -- amplifying rather than erasing the watermark signal, and (iii) orthogonal to detection -- the attack-induced bias is statistically independent of all content-side detector statistics, ensuring that amplification and evasion coexist without trade-off. Rather than perturbing generated text, SeedHijack replaces the PRNG at the supply-chain layer, biasing green-list selection without altering output tokens or degrading text quality. Across three watermarking schemes and three open-source LLMs, the attack triggers 0/6 state-of-the-art content-side statistical detectors while inflating the watermark z-score up to 2.42x (system-level defenses such as entropy-source attestation remain orthogonal and complementary). A quantum random number generator (QRNG) countermeasure is shown to fully neutralize the attack while preserving benign watermarking utility. These findings establish PRNG integrity as a first-class security requirement for cryptographic content-provenance systems.

Large Language Model (LLM) agents remain vulnerable to safety threats from the external environment, where attackers inject adversarial content into external observations such as tool-returned data, webpages, or MCP context, causing harmful agentic behaviors such as unsafe actions or incorrect outputs. Existing studies typically focus on single-interaction attacks, where the agent observes adversarial content and immediately exhibits harmful behavior within one user request. However, we show that adversarial content can also persist across interactions served by the same agent, making such threats harder to detect and mitigate. Specifically, adversarial content may persist in the agent state, remain dormant across interactions, and later be activated by a benign user query. We formalize this type of safety threat as Sleeper Attack. To evaluate it, we construct a benchmark with 1,896 instances covering six real-world harmful outcomes, three attack strategies, and three agent state targets: session context, memory, and reusable skills. Experiments on seven strong open-source and closed-source LLMs show that state-of-the-art LLM agents remain vulnerable to Sleeper Attack, even when they achieve low attack success rates under a single-interaction baseline. Our code and data are available at https://anonymous.4open.science/r/skdvnfu23ihr9wdscnksf1asdffsaef.

Retrieval-Augmented Generation (RAG) mitigates LLM hallucinations but introduces a critical vulnerability: corpus integrity. We present SilentRetrieval, a two-stage data poisoning attack that hijacks RAG systems through adversarially crafted yet fluent documents. Stage 1 uses Coordinated Beam Search, a multi-token joint optimization method with a fluency-similarity objective, to keep a poisoned host document retrievable while constraining perplexity. Stage 2 uses Context-Adaptive Trigger Generation, a lightweight trigger-fusion step driven by a frozen LLM, to integrate manipulation triggers into document content. Under a one-poisoned-document-per-query evaluation with synthetic target answers, SilentRetrieval achieves 84.6%/81.3% HR@10 and 57.5%/54.8% ASR-LLM on Natural Questions and MS MARCO, while maintaining near-benign perplexity. Cross-model evaluation across four target LLMs shows nontrivial effectiveness under a fixed trigger generator, and transfer tests against unseen retrievers, including ColBERT and commercial embedding models, yield 64.7% average HR@10 under the same injected-corpus protocol. In a sampled Wikipedia-scale evaluation, SilentRetrieval retains 74.2% HR@10 at a 0.016% poisoning ratio. Combined retrieval-side and generation-side defenses reduce attack success substantially but incur a latency trade-off. Human evaluation shows substantially lower flag rates than disfluent baselines, while remaining numerically more suspicious than benign content at the current sample size.

Linear probes trained on LLM activations are increasingly proposed as deception-detection metrics, yet report AUROC exceeding 0.96 on clean benchmarks while collapsing under distributional shift. This paper systematically pressure-tests probe-based metrics across the Gemma 3 model family (1B-27B parameters), diagnosing why they fail rather than merely documenting that they fail. We test four hypotheses about deception encoding: (1) single linear direction, (2) multi-dimensional subspace, (3) convex conic hull, (4) entropy proxy. Our design includes cross-domain transfer matrices, multi-dimensional probe analysis with permutation null baselines, entropy-residualization tests, and distractor evaluations across 8 stylistic shifts. We find that: (a) probes achieve near-perfect AUROC (>=0.998) on clean data but collapse under stylistic shifts; style-augmented probes recover near-perfect detection (mean AUROC 0.979-0.983) on unseen styles; (b) the single-direction hypothesis is rejected (k=1 captures only 0.61-0.80 AUROC), with cross-domain transfer failure confirmed as geometric rather than layer-mismatch-driven; (c) the entropy-proxy hypothesis is rejected (max |rho|=0.454, max Delta-AUROC after residualization=0.004); and (d) deception does not form a significant linear subspace (per-domain k*=0), yet multi-dimensional probes (k>=5) recover the signal through distributed sub-threshold features. Probe fragility reflects distributional narrowness rather than an architectural limitation: style-augmented probes recover near-perfect detection at both 4B and 27B, establishing that the inverse scaling pattern is a training-distribution artifact rather than a genuine scale-dependent phenomenon.

Code Large Language Models (CLLMs) serve as the core of modern code agents, enabling developers to automate complex software development tasks. In this paper, we present Poison-with-Style (PwS), a practical and stealthy model poisoning attack targeting CLLMs. Unlike prior attacks that assume an active adversary capable of directly embedding explicit triggers (e.g., specific words) into developers' prompts during inference, PwS leverages developers' code styles as covert triggers implicitly embedded within their prompts. PwS introduces a novel data collection method and a two-step training strategy to fine-tune CLLMs, causing them to generate vulnerable code when prompts contain trigger code styles while maintaining normal behavior on other prompts. Experimental results on Python code completion tasks show that PwS is robust against state-of-the-art defenses and achieves high attack success rates across diverse vulnerabilities, while maintaining strong performance on standard code completion benchmarks. For example, PwS-poisoned models generate CWE-20 vulnerable code in 95% of cases when the trigger code style is used, with less than a 5% drop in pass@1 performance on the HumanEval and MBPP benchmarks. Our implementation and dataset are here: https://github.com/khangtran2020/pws.

Large language models (LLMs) are often fine-tuned on uncurated text datasets that adversaries can poison. Existing poisoning attacks primarily rely on fixed trigger phrases that defenses such as outlier detection, clean-data regularization, or online monitoring can neutralize. In this paper, we propose a data poisoning method that teaches an LLM an information hiding scheme reliably and stealthily through semantic associations between shared knowledge such as facts or concepts and attacker-chosen phrases. The induced hiding scheme can encode and decode arbitrary malicious instructions, thus revealing a new and subtle poisoning-induced vulnerability: covert control attacks. We precisely characterize covert control attacks and evaluate them across $5$ LLMs, $3$ backdoor defenses, and $4$ prompt injection defenses. With a small poisoned fraction, covert control attacks outperform heuristic-based prompt injection attacks in average attack success rate by about $40\%$ relative to clean fine-tuned models. They also circumvent defenses based on detection and fine-tuning, maintaining up to $93\%$ attack success rate after backdoor defenses and up to $98\%$ after prompt injection defenses.

Recent defenses for safeguarding open-weight large language models (LLMs) are intended to prevent adversarial usage. Underlying these defenses is an assumption that new harmful behavior is learned through fine-tuning rather than elicited by jailbreaking the model. Yet, pretrained LLMs already encode substantial harmful knowledge across many domains, which raises an important question: can an adversary jailbreak safeguarded models, to achieve harmful usage without fine-tuning at all? In this paper, we show that open-weight safeguards are susceptible to simpler strategies that, despite being well known, have not been systematically evaluated against these safeguards. Specifically, we evaluate two low-cost attacks--abliteration and prefilling--that do not rely on gradient-based optimization. Across three harmfulness evaluation benchmarks (BeaverTails, HarmBench, and AdvBench), these attacks increase attack success rates against safeguarded open-weight models from below 10% to a range of 16%-96%. To mitigate this vulnerability, we introduce abliteration-resistant tuning (ART), which incorporates an abliteration-based objective into training. ART can be layered onto existing defenses and reduces the success rates of abliteration, prefilling, and their combination by 10%-20%. These findings indicate that the attack surface for open-weight models is broader than previously characterized, and that evaluations of safeguarding defenses should incorporate a more diverse set of attack strategies beyond adversarial fine-tuning.

While Large Language Model-based Multi-Agent Systems (LLM-MAS) demonstrate remarkable capabilities in solving complex tasks by orchestrating specialized agents and external tools, the implicit trust in tool outputs creates a critical attack surface. Existing tool attacks are limited by domain specificity or fixed and static templates. To address these challenges, we propose Evo-Attacker, which formulates the tool attack as a self-evolving, memory-augmented reinforcement learning process. Evo-Attacker constructs a dynamic attack memory and employs deliberative reasoning to retrieve adversarial patterns and strategize modifying interventions at critical moments. Furthermore, we introduce Attack-Flow GRPO to optimize intermediate reasoning steps via terminal outcomes, addressing the long-horizon credit assignment challenge. Comprehensive experiments demonstrate that Evo-Attacker consistently outperforms baselines, highlighting its generalization and evolutionary capabilities and the urgent need for defensive tool safeguards.

The known stylistic biases in LLM judges, such as a preference for verbosity or specific sentence structures, present an underexplored security vulnerability. In this work, we introduce BITE (BIas exploraTion and Exploitation), a black-box adversarial framework that learns semantics-preserving edits to mislead an LLM judge and artificially inflate the scores it assigns. We cast the selection of stylistic edits as a contextual bandit problem and use a LinUCB policy to adaptively choose edits that maximize the judge's score without access to model parameters or gradients. Empirically, we test BITE across a diverse range of LLM judges and tasks, including both pointwise and pairwise comparisons on chatbot leaderboards and AI-reviewer benchmarks. BITE achieves an attack success rate exceeding 65% and raises scores by 1-2 points on a 9-point scale, all while preserving semantic equivalence. We further assess the attack's stealthiness, showing that BITE evades standard style-control methods and several detection baselines. Our findings expose a fundamental weakness in the LLM-as-a-judge paradigm and motivate robust, attack-aware evaluation. Our code is available at https://github.com/xianglinyang/llm-as-a-judge-attack.

Large language models (LLMs) are increasingly used as analyst assistants in security operations centers (SOCs), where they ingest log and alert data to produce triage labels, incident summaries, or remediation advice. We study a structural failure mode of this design: many log fields are attacker controlled. User agents, URLs, payloads, DNS queries, and attempted usernames can therefore carry instructions to the model alongside evidence of the intrusion. We call this setting \emph{log-substrate prompt injection}. We introduce a four-class taxonomy of log-substrate attacks: direct override (S1), persona hijack (S2), context manipulation (S3), and obfuscated payloads (S4). We evaluate 48 strategy-defense-task combinations using \texttt{gpt-4o-mini} as the analyst. Three findings stand out. First, direct overrides are ineffective in our setting: all S1 classification attacks achieve 0% suppression. In contrast, persona hijacks suppress 68% of malicious logs under a naive classifier and remain effective under stronger defenses. Second, summarization is the highest-risk task: context manipulation reaches 96% injection success without defenses and 38% even with constrained output. Third, defenses reduce but do not eliminate the attack surface: average injection success falls from 26.6% under naive prompting to 11.8% under our strongest defense. We also compare empirical results to a deterministic mock analyst and find that simulation substantially mispredicts current model behavior, especially for direct overrides. These results suggest that SOC copilots should treat raw log content as adversarial input rather than ordinary analyst context.

Large language model agents increasingly rely on persistent memory to store past interactions, retrieve relevant demonstrations, and improve long-horizon task execution. However, this memory mechanism also creates a practical security vulnerability: an adversarial user may inject malicious records into the agent's memory through ordinary interaction, and these records can later be retrieved to steer the agent's reasoning and actions. Existing defenses primarily focus on online intervention, such as prompt filtering or output blocking, but they do not address the post-hoc question of which stored memories are responsible after harmful behavior has already been observed. We propose \textbf{MemAudit}, a post-hoc causal memory auditing framework for memory-augmented LLM agents. The framework combines two complementary signals: (1) a counterfactual memory influence score that measures each memory's causal contribution to harmful outputs, and (2) a memory consistency graph that identifies structurally anomalous memories within the broader memory store. We evaluate MemAudit against MINJA, a query-only memory injection attack in which malicious records are generated and stored through normal agent interactions rather than direct memory-bank modification. Across both QA and reasoning-agent settings, MemAudit substantially reduces attack success rates under realistic post-hoc auditing scenarios. The results show that QA attack success is reduced from $70\%$ to $0\%$, while RAP attack success drops from $83.3\%$ to $0\%$.

The rise of tool-using Large Language Model (LLM) agents, standardized by protocols like the Model Context Protocol (MCP), has unlocked unprecedented autonomous execution capabilities for LLM Agents by integrating external open-domain knowledge and tools. However, this interoperability introduces a covert attack surface targeting the agent's cognitive planning layer. This paper systematically investigates Tool Description Poisoning (TDP), a novel semantic attack. In TDP, malicious instructions are not embedded in a tool's executable code, but rather covertly injected into its descriptive metadata, the very "manual" an agent relies on for secure planning and decision-making. To rigorously and systematically evaluate this emerging threat, we introduce the MCP-TDP Security Benchmark. This high-fidelity sandbox environment comprises 32 realistic, real-world test cases spanning 6 distinct risk categories. Our evaluation of 8 mainstream LLMs reveals severe vulnerabilities, with leading models like GPT-4o exhibiting a nearly 100% Attack Success Rate (ASR) in six high-risk scenarios. Furthermore, our findings demonstrate that common prompt-guardrail defenses are largely ineffective and can, counterintuitively, even be counterproductive (a phenomenon which we term the "Firewall Fallacy"). Crucially, we also propose a defense mechanism: "Reactive Self-Correction," where an agent autonomously detects and reverts its own malicious actions post-execution. This work provides the first specialized security benchmark tailored for TDP, offering essential insights for securing the cognitive and planning layers of advanced agentic systems.

When practitioners fine-tune LLMs on unvetted datasets, an adversary can exploit the data supply chain through task-level poisoning: inserting a small number of crafted instruction-response pairs that cause the model to embed attacker-specified entities, such as a country, in outputs for a targeted task family while behaving normally elsewhere. We introduce PoisonForge, a benchmark that parameterizes this threat along four dimensions (bias type, poisoning mode, appearance count, and target output length) and evaluates 12 open-weight models (from 2B to 32B parameters) across five families under a primarily 1% poison budget. With only 10 poisoned examples among 1,000 fine-tuning examples, 11 of 12 models exceed a 70% attack success rate (ASR) in their most vulnerable configuration. Meanwhile, unintended leakage to non-target tasks remains below 0.5%, and models perform well on standard benchmarks. We analyze in detail the factors contributing to attack success. We observe that multiple appearances of an entity increase the ASR, the optimal poisoning mode depends on the semantic structure of the target entity, and ASR drops monotonically with the task output length. A correlation analysis and risk prediction model confirm that poisoning design choices, rather than model scale, are the primary causes of attack success, and that these patterns generalize to predict attack success on new tasks. We release all configurations, pipelines, and analysis code to support reproducible comparisons.

A large language model (LLM) agent that sends messages can leak data inside them. Destination allowlists and content scanners do not police whether an otherwise-benign payload is itself a covert channel: a compromised agent encodes bits in zero-width characters, homoglyphs, whitespace, base64, JavaScript Object Notation (JSON) key ordering, message timing or size -- and, in binary egress, in least-significant-bit (LSB) pixel planes, per-image mean luminance, inter-image sequence permutation, ultrasonic tones, or audible-band sonified data. Our egress reference monitor has three contributions. (i) A text pipeline of ten capacity-reducing stages, a per-sink leaky-bucket capacity ledger, and a staged posture that enforces lossless stages from day one. (ii) Two media scramblers (a Fourier-domain audio band-limiter and a red-green-blue (RGB) image bit-depth and mean-luminance bucketer) gated by a boot-time cryptographic legitimacy attestation: an auditor publishes at boot the trusted Ed25519 keys and {kind, data-class} pairs; only payloads with a verifying signature for an authorized class are exempt. The attestation sidesteps the intractable content-based discrimination between real media and data sonified or rasterized as a carrier; unsigned media is suspect by default; a content-addressed canonicalizer closes the inter-image permutation channel. (iii) Residual capacity is the Miller--Madow corrected mutual information between embedded and recovered bits (zero when destroyed), measured by an adversarial ensemble of fifteen working encoders across text, image and audio. The reference implementation drives residual capacity to zero on every destroyable channel and to a stated bound on the one (per-image mean luminance) that cannot be destroyed without ruining the image.