Mind the Gap: Evaluating LLMs for High-Level Malicious Package Detection Vs. Fine-Grained Indicator Identification

The prevalence of malicious packages in open-source repositories, such as PyPI, poses a critical threat to the software supply chain. While Large Language Models LLMs have emerged as a promising tool for automated security tasks, their effectiveness in detecting malicious packages and indicators...

NLTK 安全漏洞

NLTK is an open-source natural language toolkit developed by NLTK. It is used to support research and development in natural language processing. NLTK has a security vulnerability that stems from the unzipiter function using zipfile.extractall without performing path validation or security checks...

Splunk Cloud Platform和Splunk Enterprise 日志信息泄露漏洞

Splunk Cloud Platform and Splunk Enterprise are both products of the American company Splunk. Splunk Cloud Platform is a powerful service for data collection, processing, and analysis. Splunk Enterprise is a suite of software for data collection and analysis. Both Splunk Enterprise and Splunk Clo...

ezBookkeeping 安全漏洞

ezBookkeeping is a lightweight personal accounting application developed by mayswind developers. Versions of ezBookkeeping 1.2.0 and earlier contained security vulnerabilities. These vulnerabilities stemmed from the lack of validation of nested depths during the processing of JSON and XML file...

Ctags 安全漏洞

CTags is a utility developed by SourceForge OpenSource, designed to generate index files from the program source code tree, thereby facilitating quick location of text in text editors. Ctags versions 6.2.1 and earlier contained security vulnerabilities, which stemmed from improper recursive contr...

PT-2026-20389

When hours are entered in time@work, version 7.0.5, it performs a query to display the projects assigned to the user. If the query URL is copied and opened in a new browser window, the ‘IDClient’ parameter is vulnerable to a blind authenticated SQL injection. If the request is made with the TWAdm...

SQUIRREL 安全漏洞

SQUIRREL is a programming language developed by Alberto Demichelis. It is the stable version of SQUIRREL 3.2. Versions of SQUIRREL 3.2 and earlier have security vulnerabilities. These vulnerabilities stem from a heap buffer overflow in the SQObjectPtr::operator function in the sqobject.h library,...

PT-2026-20354

A weakness has been identified in universal-ctags ctags up to 6.2.1. The affected element is the function parseExpression/parseExprList of the file parsers/v.c of the component V Language Parser. Executing a manipulation can lead to uncontrolled recursion. It is possible to launch the attack on t...

XHan Admin SQL注入漏洞

XHan Admin is a management system developed by Alixhan’s individual developers. Versions of XHan Admin prior to 1.7.0 contained an SQL injection vulnerability. This vulnerability stemmed from incorrect handling of parameters in files/frontend-api/system-service/api/system/role/query, specifically...

lily 缓冲区错误漏洞

Lily is a programming language developed by FascinatedBox’s individual developers. Versions of Lily prior to 2.3 contained a buffer error vulnerability. This vulnerability stems from an out-of-bounds read in the counttransforms function located in the src/lilyemitter.c file, which could lead to...

Discovering Universal Activation Directions for PII Leakage in Language Models

Modern language models exhibit rich internal structure, yet little is known about how privacy-sensitive behaviors, such as personally identifiable information PII leakage, are represented and modulated within their hidden states. We present UniLeak, a mechanistic-interpretability framework that...

SQUIRREL 缓冲区错误漏洞

SQUIRREL is the stable version of the programming language SQUIRREL 3.2. Versions of SQUIRREL 3.2 and earlier had a buffer error vulnerability, which stemmed from improper handling of parameters and the target stack. This vulnerability could lead to out-of-bounds read operations...

wren 安全漏洞

Wren is a concurrent scripting language developed by Wren OpenSource. Versions of Wren prior to 0.4.0 have security vulnerabilities; these vulnerabilities stem from stack buffer overflows in the error message processing component, which may lead to local memory corruption...

Delinea Cloud Suite 安全漏洞

Delinea Cloud Suite is a cloud-based resource pool management software developed by Delinea Corporation in the United States. Delinea Cloud Suite has a security vulnerability that stems from improper handling of special elements within SQL commands, which may lead to SQL injection attacks...

Linux Distros Unpatched Vulnerability : CVE-2026-2641

The Linux/Unix host has one or more packages installed that are impacted by a vulnerability without a vendor supplied patch available. - A weakness has been identified in universal-ctags ctags up to 6.2.1. The affected element is the function parseExpression/parseExprList of the file parsers/v.c ...

BSV Blockchain SDK has an Authentication Signature Data Preparation Vulnerability

BRC-104 Authentication Signature Data Preparation Vulnerability Summary A critical cryptographic vulnerability in the TypeScript SDK's BRC-104 authentication implementation caused incorrect signature data preparation, resulting in signature incompatibility between SDK implementations and potentia...

GHSA-VJPQ-XX5G-QVMM BSV Blockchain SDK has an Authentication Signature Data Preparation Vulnerability

BRC-104 Authentication Signature Data Preparation Vulnerability Summary A critical cryptographic vulnerability in the TypeScript SDK's BRC-104 authentication implementation caused incorrect signature data preparation, resulting in signature incompatibility between SDK implementations and potentia...

CVE-2025-8303

Improper Neutralization of Input During Web Page Generation XSS or 'Cross-site Scripting' vulnerability in EKA Software Computer Information Advertising Services Ltd. Real Estate Script V5 With Doping Module – Store Module – New Language System allows Cross-Site Scripting XSS. This issue affects...

Side-Channel Attacks Against LLMs

Here are three papers describing different side-channel attacks against LLMs. "Remote Timing Attacks on Efficient Language Model Inference": Abstract: Scaling up language models has significantly increased their capabilities. But larger models are slower models, and so there is now an extensive...

CVE-2025-8303

Improper Neutralization of Input During Web Page Generation XSS or 'Cross-site Scripting' vulnerability in EKA Software Computer Information Advertising Services Ltd. Real Estate Script V5 With Doping Module – Store Module – New Language System allows Cross-Site Scripting XSS. This issue affects...