Exploiting Vulnerabilities: Universal Adversarial Attacks on Vision-Language-Action Models in Robotics

Recently, Vision-Language-Action VLA models have revolutionized robotic manipulation by seamlessly integrating visual perception, language understanding, and action generation in an end-to-end learning framework. However, since these models are designed to interact directly with the physical worl...

Do Defenses against LLM Extraction Work across Attacks? A Lifecycle Benchmark of Black-Box Model Extraction

Large language models LLMs deployed through text-only APIs face model extraction risks, as adversaries can collect their responses to train surrogates that reproduce their capabilities. While prior work has developed diverse attacks and defenses, evaluations remain fragmented across access...

CVE-2026-93764

A flaw was found in Mongoid. When generating the client-side field-level encryption schema, Mongoid may fail to apply encryption rules to fields within embedded models. This oversight can lead to sensitive data, intended for encryption, being stored in plaintext without any indication or error...

Security-Enhanced Seed-Based Weight Quantization for Large Language Models

Large language models LLMs incur substantial storage, memory-bandwidth and energy costs, motivating compact weight representations. Existing seed-based compression methods reconstruct weights from compact pseudo-random representations but do not explicitly account for the non-uniform sensitivity ...

CollageAttack: Exploiting Cross-Modal Alignment Flaws in T2I Models through Spatial Text Composition

Text-to-image T2I models have substantially improved in language understanding, in-image text rendering, and visual composition, while their safety mechanisms do not always keep pace with these capabilities. This creates a cross-modal attack surface in which harmful semantics can remain...

Controlled Decoding Attacks on Black-Box LLMs

Manipulating next-token probabilities during generation can bypass the safety alignment of large language models. Existing approaches, however, rely on access to model weights or numerical token probabilities and therefore do not apply to interfaces that return only sampled text. Reconstructing...

Removing the NEEDLE in the Haystack: Backdoor Removal in LLMs Via Weight Orthogonalisation

Backdoor attacks can be implanted in Large Language Models LLMs during training, causing unwanted behaviour when a trigger appears in the input. Existing backdoor defences for LLMs attempt to remove the backdoor but inadvertently shift the model's output distribution to benign prompts, which can...

Where Do LLMs Decide to Break the Rules? Mechanistic Localization of Prompt Injection Compliance

When a prompt injection attack succeeds, a Large Language Model LLM abandons its assigned system role to comply with an adversarial instruction. While prior work has extensively quantified how often this occurs, we ask a more fundamental question: where inside the network does the model actually...

Selective Channel Restoration for Backdoored Vision-Language Models

Vision-language models VLMs exhibit strong multimodal capabilities but remain vulnerable to backdoors implanted through poisoned fine-tuning data. Existing defenses often require extensive parameter updates during fine-tuning or incur per-query overhead during inference. To address these...

Deep Learning Latency Attacks and Defenses: A Cross-Domain Survey of Availability Threats

Adversarial machine learning has focused mainly on integrity, but availability is an increasingly consequential complement. Latency attacks also energy-latency attacks increase inference-time work, energy, or response time, causing deadline misses, throughput collapse, or resource exhaustion in...

The Geometry of Harmfulness in Multi-Turn Attacks

Large language models LLMs remain vulnerable to adversarial attacks that circumvent safety alignment to elicit harmful outputs. It remains unclear how harmfulness and refusal representations evolve over the course of multi-turn attacks, and why single-turn defenses are less effective in multi-tur...

Distillation Defenses Easily Break after Reinforcement Learning

Distillation attacks copy the reasoning capabilities of closed-source large language models, allowing bad actors to replicate state-of-the-art performance at low cost. Attackers systematically collect a large volume of frontier model reasoning traces and then train i.e., "distill" their own model...

CVE-2026-101041

The account recovery password reset functionality in the vulnerability-lookup web application contains a time-of-check-to-time-of-use TOCTOU race condition in the consumption of single-use recovery tokens. The original implementation verified the token nonce against the stored digest and then...

CVE-2026-101041 Vulnerability-Lookup - Race Condition in Account Recovery Token Consumption Allows Password Takeover

The account recovery password reset functionality in the vulnerability-lookup web application contains a time-of-check-to-time-of-use TOCTOU race condition in the consumption of single-use recovery tokens. The original implementation verified the token nonce against the stored digest and then...

CVE-2026-101041 Vulnerability-Lookup - Race Condition in Account Recovery Token Consumption Allows Password Takeover

The account recovery password reset functionality in the vulnerability-lookup web application contains a time-of-check-to-time-of-use TOCTOU race condition in the consumption of single-use recovery tokens. The original implementation verified the token nonce against the stored digest and then...

CVE-2026-101041 Vulnerability-Lookup - Race Condition in Account Recovery Token Consumption Allows Password Takeover

The account recovery password reset functionality in the vulnerability-lookup web application contains a time-of-check-to-time-of-use TOCTOU race condition in the consumption of single-use recovery tokens. The original implementation verified the token nonce against the stored digest and then...

Evaluating System One Models for Agent Security Decisions: Reliability, Calibration, and Selective Automation

Model-based judges support agent security by detecting prompt injections, assessing interaction risks, and screening harmful requests. System One models expose typed decisions with probabilities that software can use to allow, block, or escalate inputs, but whether these probabilities support...

No Free Efficiency: Revisiting the Trade-Off between Training Efficiency and Model Vulnerability

Training efficiency has become the central driver of recent progress in foundation models. To overcome the massive computational and data requirements of large-scale training, researchers increasingly adopt strategies such as selective data sampling, efficient pre-training, and simplified...

PT-2026-99549

Name of the Vulnerable Software and Affected Versions vulnerability-lookup affected versions not specified Description The account recovery functionality contains a time-of-check-to-time-of-use TOCTOU race condition—a scenario where a system checks a condition and then uses the result, but the...

Towards Understanding LLM-Based Log Anomaly Detection: An Empirical Study of Performance, Efficiency, and Robustness

Large language models LLMs have demonstrated promising performance in log anomaly detection, yet how their adaptation strategies, architectures, and deployment configurations affect detection effectiveness remains insufficiently understood. To investigate these factors, we conduct a systematic...