Chain-Of-Code Collapse: Reasoning Failures in LLMs Via Adversarial Prompting in Code Generation

Large Language Models LLMs have achieved remarkable success in tasks requiring complex reasoning, such as code generation, mathematical problem solving, and algorithmic synthesis -- especially when aided by reasoning tokens and Chain-of-Thought prompting. Yet, a core question remains: do these...

SoK: Evaluating Jailbreak Guardrails for Large Language Models

Large Language Models LLMs have achieved remarkable progress, but their deployment has exposed critical vulnerabilities, particularly to jailbreak attacks that circumvent safety mechanisms. Guardrails--external defense mechanisms that monitor and control LLM interaction--have emerged as a promisi...

LLMs Cannot Reliably Judge (Yet?): a Comprehensive Assessment on the Robustness of LLM-As-A-Judge

Large Language Models LLMs have demonstrated remarkable intelligence across various tasks, which has inspired the development and widespread adoption of LLM-as-a-Judge systems for automated model testing, such as red teaming and benchmarking. However, these systems are susceptible to adversarial...

Prompt Attacks Reveal Superficial Knowledge Removal in Unlearning Methods

In this work, we show that some machine unlearning methods may fail when subjected to straightforward prompt attacks. We systematically evaluate eight unlearning techniques across three model families, and employ output-based, logit-based, and probe analysis to determine to what extent supposedly...

EulerOS 2.0 SP12 : libxslt (EulerOS-SA-2025-1600)

According to the versions of the libxslt package installed, the EulerOS installation on the remote host is affected by the following vulnerabilities : numbers.c in libxslt before 1.1.43 has a use-after-free because, in nested XPath evaluations, an XPath context node can be modified but never...

SinoTrack GPS Receiver

RISK EVALUATION Successful exploitation of these vulnerabilities could allow an attacker to access device profiles for which they are not authorized through the common web management interface. Access to the device profile may allow an attacker to perform some remote functions on connected...

SoK: Machine Unlearning for Large Language Models

Large language model LLM unlearning has become a critical topic in machine learning, aiming to eliminate the influence of specific training data or knowledge without retraining the model from scratch. A variety of techniques have been proposed, including Gradient Ascent, model editing, and...

Boosting Gradient Leakage Attacks: Data Reconstruction in Realistic FL Settings

Federated learning FL enables collaborative model training among multiple clients without the need to expose raw data. Its ability to safeguard privacy, at the heart of FL, has recently been a hot-button debate topic. To elaborate, several studies have introduced a type of attacks known as gradie...

Your Agent Can Defend Itself against Backdoor Attacks

Despite their growing adoption across domains, large language model LLM-powered agents face significant security risks from backdoor attacks during training and fine-tuning. These compromised agents can subsequently be manipulated to execute malicious operations when presented with specific...

GHSA-VQVV-2WJ5-Q34W Authorino Uncontrolled Resource Consumption vulnerability

A Developer persona can bring down the Authorino service, preventing the evaluation of all AuthPolicies on the cluster...

Authorino Uncontrolled Resource Consumption vulnerability

A Developer persona can bring down the Authorino service, preventing the evaluation of all AuthPolicies on the cluster...

CVE-2025-25208

A Developer persona can bring down the Authorino service, preventing the evaluation of all AuthPolicies on the cluster...

CVE-2025-25208 Rhcl: authorino denial of service through authpolicy with sharedsecretref severity

A Developer persona can bring down the Authorino service, preventing the evaluation of all AuthPolicies on the cluster...

PT-2025-24400 · Authorino · Authorino

Name of the Vulnerable Software and Affected Versions: Authorino affected versions not specified Description: A Developer persona can bring down the Authorino service, preventing the evaluation of all AuthPolicies on the cluster. Recommendations: At the moment, there is no information about a new...

Secure Distributed Learning for CAVs: Defending against Gradient Leakage with Leveled Homomorphic Encryption

Federated Learning FL enables collaborative model training across distributed clients without sharing raw data, making it a promising approach for privacy-preserving machine learning in domains like Connected and Autonomous Vehicles CAVs. However, recent studies have shown that exchanged model...

SoK: Data Reconstruction Attacks against Machine Learning Models: Definition, Metrics, and Benchmark

Data reconstruction attacks, which aim to recover the training dataset of a target model with limited access, have gained increasing attention in recent years. However, there is currently no consensus on a formal definition of data reconstruction attacks or appropriate evaluation metrics for...

MalGEN: a Generative Agent Framework for Modeling Malicious Software in Cybersecurity

The dual use nature of Large Language Models LLMs presents a growing challenge in cybersecurity. While LLM enhances automation and reasoning for defenders, they also introduce new risks, particularly their potential to be misused for generating evasive, AI crafted malware. Despite this emerging...

Can In-Context Reinforcement Learning Recover from Reward Poisoning Attacks?

We study the corruption-robustness of in-context reinforcement learning ICRL, focusing on the Decision-Pretrained Transformer DPT, Lee et al., 2023. To address the challenge of reward poisoning attacks targeting the DPT, we propose a novel adversarial training framework, called Adversarially...

SpiceDB 安全特征问题漏洞

SpiceDB is a fine-grained permissions database from the Authzed team. A security feature issue vulnerability exists in SpiceDB versions prior to 1.44.2, which stems from a schema evaluation issue involving arrow relationships that could result in an incorrect response...

Benchmarking Misuse Mitigation against Covert Adversaries

Existing language model safety evaluations focus on overt attacks and low-stakes tasks. Realistic attackers can subvert current safeguards by requesting help on small, benign-seeming tasks across many independent queries. Because individual queries do not appear harmful, the attack is hard to...