Hash Collision Attack

vllm is vulnerable to hash collision and data integrity issues. The vulnerability is due to improper image serialization using only raw pixel bytes without metadata, allowing attackers to create images with identical hashes and exploit cache poisoning or access sensitive data...

Trojan Horse Hunt in Time Series Forecasting for Space Operations

This competition hosted on Kaggle https://www.kaggle.com/competitions/trojan-horse-hunt-in-space is the first part of a series of follow-up competitions and hackathons related to the "Assurance for Space Domain AI Applications" project funded by the European Space Agency...

Which Factors Make Code LLMs More Vulnerable to Backdoor Attacks? A Systematic Study

Code LLMs are increasingly employed in software development. However, studies have shown that they are vulnerable to backdoor attacks: when a trigger a specific input pattern appears in the input, the backdoor will be activated and cause the model to generate malicious outputs. Researchers have...

Exploit for HTTP Request Smuggling in Apache Http_Server

CVE 2023 25690 - Proof of Concept Published: 7 March 2023...

Security Concerns for Large Language Models: a Survey

Large Language Models LLMs such as GPT-4 and its recent iterations, Google's Gemini, Anthropic's Claude 3 models, and xAI's Grok have caused a revolution in natural language processing, but their capabilities also introduce new security vulnerabilities. In this survey, we provide a comprehensive...

Towards Secure MLOps: Surveying Attacks, Mitigation Strategies, and Research Challenges

The rapid adoption of machine learning ML technologies has driven organizations across diverse sectors to seek efficient and reliable methods to accelerate model development-to-deployment. Machine Learning Operations MLOps has emerged as an integrative approach addressing these requirements by...

Adversarial Threat Vectors and Risk Mitigation for Retrieval-Augmented Generation Systems

Retrieval-Augmented Generation RAG systems, which integrate Large Language Models LLMs with external knowledge sources, are vulnerable to a range of adversarial attack vectors. This paper examines the importance of RAG systems through recent industry adoption trends and identifies the prominent...

Cybercriminals Target AI Users with Malware-Loaded Installers Posing as Popular Tools

Fake installers for popular artificial intelligence AI tools like OpenAI ChatGPT and InVideo AI are being used as lures to propagate various threats, such as the CyberLock and LuckyGh0$t ransomware families, and a new malware dubbed Numero. "CyberLock ransomware, developed using PowerShell,...

Spa-VLM: Stealthy Poisoning Attacks on RAG-Based VLM

With the rapid development of the Vision-Language Model VLM, significant progress has been made in Visual Question Answering VQA tasks. However, existing VLM often generate inaccurate answers due to a lack of up-to-date knowledge. To address this issue, recent research has introduced...

Employees Searching Payroll Portals on Google Tricked Into Sending Paychecks to Hackers

Threat hunters have exposed a novel campaign that makes use of search engine optimization SEO poisoning techniques to target employee mobile devices and facilitate payroll fraud. The activity, first detected by ReliaQuest in May 2025 targeting an unnamed customer in the manufacturing sector, is...

ABB M2M Gateway HTTP Request Smuggling in embedded Apache HTTP Server (CVE-2023-25690)

Some modproxy configurations on Apache HTTP Server versions 2.4.0 through 2.4.55 allow a HTTP Request Smuggling attack. Configurations are affected when modproxy is enabled along with some form of RewriteRule or ProxyPassMatch in which a non-specific pattern matches some portion of the...

ABB M2M Gateway HTTP Request Smuggling in embedded Bind (CVE-2021-25220)

BIND 9.11.0 - 9.11.36 9.12.0 - 9.16.26 9.17.0 - 9.18.0 BIND Supported Preview Editions: 9.11.4-S1 - 9.11.36-S1 9.16.8-S1 - 9.16.26-S1 Versions of BIND 9 earlier than those shown - back to 9.1.0, including Supported Preview Editions - are also believed to be affected but have not been tested as th...

CPA-RAG:Covert Poisoning Attacks on Retrieval-Augmented Generation in Large Language Models

Retrieval-Augmented Generation RAG enhances large language models LLMs by incorporating external knowledge, but its openness introduces vulnerabilities that can be exploited by poisoning attacks. Existing poisoning methods for RAG systems have limitations, such as poor generalization and lack of...

Benchmarking Poisoning Attacks against Retrieval-Augmented Generation

Retrieval-Augmented Generation RAG has proven effective in mitigating hallucinations in large language models by incorporating external knowledge during inference. However, this integration introduces new security vulnerabilities, particularly to poisoning attacks. Although prior work has explore...

CVE-2025-20255

A vulnerability in client join services of Cisco Webex Meetings could allow an unauthenticated, remote attacker to manipulate cached HTTP responses within the meeting join service. This vulnerability is due to improper handling of malicious HTTP requests to the affected service. An attacker could...

CVE-2024-27185

The pagination class includes arbitrary parameters in links, leading to cache poisoning attack vectors...

CVE-2024-6451

AI Engine 2.4.3 is susceptible to remote-code-execution RCE via Log Poisoning. The AI Engine WordPress plugin before 2.5.1 fails to validate the file extension of "logspath", allowing Administrators to change log filetypes from .log to .php...

CVE-2024-1554

The fetch API and navigation incorrectly shared the same cache, as the cache key did not include the optional headers fetch may contain. Under the correct circumstances, an attacker may have been able to poison the local browser cache by priming it with a fetch response controlled by the addition...

CVE-2024-29042

Translate is a package that allows users to convert text to different languages on Node.js and the browser. Prior to version 3.0.0, an attacker controlling the second variable of the translate function is able to perform a cache poisoning attack. They can change the outcome of translation request...

CVE-2024-54143

openwrt/asu is an image on demand server for OpenWrt based distributions. The request hashing mechanism truncates SHA-256 hashes to only 12 characters. This significantly reduces entropy, making it feasible for an attacker to generate collisions. By exploiting this, a previously built malicious...