Robust Anti-Backdoor Instruction Tuning in LVLMs

Large visual language models LVLMs have demonstrated excellent instruction-following capabilities, yet remain vulnerable to stealthy backdoor attacks when finetuned using contaminated data. Existing backdoor defense techniques are usually developed for single-modal visual or language models under...

CVE-2024-3509 Stored Cross-Site Scripting (XSS) in Management Console of Multiple WSO2 Products via Rich Text Editor

A stored cross-site scripting XSS vulnerability exists in the Management Console of multiple WSO2 products due to insufficient input validation in the Rich Text Editor within the registry section. To exploit this vulnerability, a malicious actor must have a valid user account with administrative...

CVE-2024-3509

CVE-2024-3509 is a stored XSS in the Management Console of multiple WSO2 products caused by insufficient input validation in the Rich Text Editor within the registry section. Exploitation requires a valid admin account; a successful attack can inject persistent JavaScript payloads to exfiltrate d...

CVE-2024-3509 Stored Cross-Site Scripting (XSS) in Management Console of Multiple WSO2 Products via Rich Text Editor

A stored cross-site scripting XSS vulnerability exists in the Management Console of multiple WSO2 products due to insufficient input validation in the Rich Text Editor within the registry section. To exploit this vulnerability, a malicious actor must have a valid user account with administrative...

CVE-2025-5447

A vulnerability was found in Linksys RE6500, RE6250, RE6300, RE6350, RE7000 and RE9000 1.0.013.001/1.0.04.001/1.0.04.002/1.1.05.003/1.2.07.001. It has been declared as critical. This vulnerability affects the function ssid1MACFilter of the file /goform/ssid1MACFilter. The manipulation of the...

Align Is Not Enough: Multimodal Universal Jailbreak Attack against Multimodal Large Language Models

Large Language Models LLMs have evolved into Multimodal Large Language Models MLLMs, significantly enhancing their capabilities by integrating visual information and other types, thus aligning more closely with the nature of human intelligence, which processes a variety of data forms beyond just...

Dot 跨站脚本漏洞

Dot is a text-to-speech, RAG and LLM tool by alexpinel individual developers. A cross-site scripting vulnerability exists in Dot 0.9.3 and earlier versions, which stems from user input and LLM output being appended to the DOM using innerHTML, which could lead to cross-site scripting and command...

Privacy-Aware, Public-Aligned: Embedding Risk Detection and Public Values into Scalable Clinical Text De-Identification for Trusted Research Environments

Clinical free-text data offers immense potential to improve population health research such as richer phenotyping, symptom tracking, and contextual understanding of patient care. However, these data present significant privacy risks due to the presence of directly or indirectly identifying...

Security update for postgresql14

This update for postgresql14 fixes the following issues: Upgrade to 14.18: CVE-2025-4207: Fixed PostgreSQL GB18030 encoding validation can read one byte past end of allocation for text that fails validation bsc1242931 Patch Instructions: To install this SUSE update use the SUSE recommended...

Security update for postgresql17

This update for postgresql17 fixes the following issues: Upgrade to 17.5: CVE-2025-4207: Fixed PostgreSQL GB18030 encoding validation can read one byte past end of allocation for text that fails validation bsc1242931 Changelog: https://www.postgresql.org/docs/release/17.5/ Patch Instructions: To...

[SECURITY] Fedora 42 Update: coreutils-9.6-4.fc42

These are the GNU core utilities. This package is the combination of the old GNU fileutils, sh-utils, and textutils packages...

Tinxy WiFi Lock Controller 安全漏洞

Tinxy WiFi Lock Controller is a smart door lock from Tinxy. A security vulnerability exists in Tinxy WiFi Lock Controller v1. The vulnerability stems from storing sensitive user information in clear text, which may lead to information leakage...

Tinxy WiFi Lock Controller v1 RF 安全漏洞

Tinxy WiFi Lock Controller v1 RF is a smart door lock from Tinxy. A security vulnerability exists in Tinxy WiFi Lock Controller v1 RF, which originates from transmitting sensitive information in clear text and could lead to a man-in-the-middle attack...

Security update for postgresql13

This update for postgresql13 fixes the following issues: Upgrade to 13.21: CVE-2025-4207: Fixed PostgreSQL GB18030 encoding validation can read one byte past end of allocation for text that fails validation bsc1242931 Changelog: https://www.postgresql.org/docs/release/13.21/ Patch Instructions: T...

Fooling the Watchers: Breaking AIGC Detectors Via Semantic Prompt Attacks

The rise of text-to-image T2I models has enabled the synthesis of photorealistic human portraits, raising serious concerns about identity misuse and the robustness of AIGC detectors. In this work, we propose an automated adversarial prompt generation framework that leverages a grammar tree...

Important: thunderbird

Issue Overview: Through a series of popup and window.print calls, an attacker can cause a window to go fullscreen without the user seeing the notification prompt, resulting in potential user confusion or spoofing attacks. This vulnerability affects Firefox ESR 102.5, Thunderbird 102.5, and Firefo...

New warning issued over toll fee scams

Over a year ago the FBI warned about what was then a new form of smishing phishing via SMS scam: text messages that demanded payment for toll fees. The FTC sent out a similar warning in January, 2025. Then, in April another wave of toll fee scams began doing the rounds. Now the Departments of Mot...

vuetify 跨站脚本漏洞

vuetify is a material component framework for Vue open source by vuetify Germany. A cross-site scripting vulnerability exists in versions prior to vuetify 3.0.0, which stems from the eventMoreText property of the VCalendar component not being cleaned up correctly, which could lead to a cross-site...

The vulnerability of Siemens Scalance LPE9403 industrial switches’ microprogramming software, related to the transmission of secret information in the form of open text, allows a intruder to gain unauthorized access to the protected information.

The vulnerability of Siemens Scalance LPE9403 industrial switches’ microprogramming software relates to the transmission of secret information in the form of open text. Exploiting this vulnerability can allow an intruder to gain unauthorized access to the protected information...

The Feasibility of Topic-Based Watermarking on Academic Peer Reviews

Large language models LLMs are increasingly integrated into academic workflows, with many conferences and journals permitting their use for tasks such as language refinement and literature summarization. However, their use in peer review remains prohibited due to concerns around confidentiality...