dify 安全漏洞

dify is an open-source LLM application development platform developed by LangGenius. Versions of dify prior to 1.9.0 contained security vulnerabilities. These vulnerabilities were caused by differences in API responses, which could lead to the enumeration of registered email addresses...

Reverse CAPTCHA: Evaluating LLM Susceptibility to Invisible Unicode Instruction Injection

We introduce Reverse CAPTCHA, an evaluation framework that tests whether large language models follow invisible Unicode-encoded instructions embedded in otherwise normal-looking text. Unlike traditional CAPTCHAs that distinguish humans from machines, our benchmark exploits a capability gap: model...

CVE-2026-26744

A user enumeration vulnerability exists in FormaLMS 4.1.18 and below in the password recovery functionality accessible via the /lostpwd endpoint. The application returns different error messages for valid and invalid usernames allowing an unauthenticated attacker to determine which usernames are...

CVE-2026-26744

A user enumeration vulnerability exists in FormaLMS 4.1.18 and below in the password recovery functionality accessible via the /lostpwd endpoint. The application returns different error messages for valid and invalid usernames allowing an unauthenticated attacker to determine which usernames are...

Apache Shiro 安全漏洞

Apache Shiro is a Java security framework developed by the Apache Foundation in the United States. It is used for authentication, authorization, encryption, and session management. Versions of Apache Shiro such as 1. and 2.0.7 had security vulnerabilities. These vulnerabilities were due to observ...

CVE-2025-61732 Potential code smuggling via doc comments in cmd/cgo

A discrepancy between how Go and C/C++ comments were parsed allowed for code smuggling into the resulting cgo binary...

CVE-2026-24664

The Open eClass platform formerly known as GUnet eClass is a complete course management system. Prior to version 4.2, a username enumeration vulnerability allows unauthenticated attackers to identify valid user accounts by analyzing differences in the login response behavior. This issue has been...

CVE-2026-24664 Open eClass is Vulnerable to Username Enumeration via Login Response Discrepancies

The Open eClass platform formerly known as GUnet eClass is a complete course management system. Prior to version 4.2, a username enumeration vulnerability allows unauthenticated attackers to identify valid user accounts by analyzing differences in the login response behavior. This issue has been...

PT-2026-28865

Уязвимость программной платформы для веб-приложений Django связана с раскрытием информации на основании временных расхождений. Эксплуатация уязвимости может позволить нарушителю, действующему удаленно, получить несанкционированный доступ к защищаемой информации...

vLLM code issues and vulnerabilities

vLLM is an open-source solution designed for LLM-based models, featuring high throughput and memory-efficient reasoning and service engines. Prior to vLLM 0.14.1, there were code-related vulnerabilities. These vulnerabilities stemmed from differences in the interpretation of backslashes by variou...

CVE-2025-22234

CVE-2025-22234 is associated with Spring Security’s timing-attack mitigation in DaoAuthenticationProvider. The described issue states that the fix applied in CVE-2025-22228 accidentally broke the mitigation, enabling an attacker to infer usernames or authentication behavior via response-time diff...

SUSE CVE-2025-69413

In Gitea before 1.25.2, /api/v1/user has different responses for failed authentication depending on whether a username exists...

V-APF

V-APF 是一款机器学习驱动的自动化 Web 安全检测框架。通过对 Web 页面在“基线请求 vs. 探测请求”间的语义指纹差异进行量化，结合随机森林 Random Forest模型，自动识别 SQLi / XSS 等漏洞并生成报告。...

A Longitudinal Measurement Study of Log4Shell Exploitation from an Active Network Telescope

The disclosure of the Log4Shell vulnerability in December 2021 led to an unprecedented wave of global scanning and exploitation activity. A recent study provided important initial insights, but was largely limited in duration and geography, focusing primarily on European and U.S. network telescop...

CVE-2024-55374

REDCap 14.3.13 allows an attacker to enumerate usernames due to an observable discrepancy between login attempts...

REDCap 安全漏洞

REDCap is a data collection and management web application from REDCap Open Source. A security vulnerability exists in REDCap version 14.3.13 that stems from an observable difference between login attempts and could lead an attacker to enumerate usernames...

Emoji-Based Jailbreaking of Large Language Models

Large Language Models LLMs are integral to modern AI applications, but their safety alignment mechanisms can be bypassed through adversarial prompt engineering. This study investigates emoji-based jailbreaking, where emoji sequences are embedded in textual prompts to trigger harmful and unethical...

CVE-2024-55374

REDCap version 14.3.13 is vulnerable to username enumeration because the application returns different error messages based on whether a user exists during login attempts. An attacker can exploit this observable discrepancy by sending multiple HTTP authentication requests to perform brute-force a...

CVE-2023-53943

GLPI 9.5.7 contains a username enumeration vulnerability in the lost password recovery mechanism that allows attackers to validate email addresses. Attackers can systematically test email addresses by submitting requests to the password reset endpoint and analyzing response differences to identif...

UBUNTU-CVE-2025-66567

The ruby-saml library is for implementing the client side of a SAML authorization. ruby-saml versions up to and including 1.12.4 contain an authentication bypass vulnerability due to an incomplete fix for CVE-2025-25292. ReXML and Nokogiri parse XML differently, generating entirely different...