Cybersecurity AI: Evaluating Agentic Cybersecurity in Attack/Defense CTFs

We empirically evaluate whether AI systems are more effective at attacking or defending in cybersecurity. Using CAI Cybersecurity AI's parallel execution framework, we deployed autonomous agents in 23 Attack/Defense CTF battlegrounds. Statistical analysis reveals defensive agents achieve 54.3%...

Multimodal Safety Is Asymmetric: Cross-Modal Exploits Unlock Black-Box MLLMs Jailbreaks

Multimodal large language models MLLMs have demonstrated significant utility across diverse real-world applications. But MLLMs remain vulnerable to jailbreaks, where adversarial inputs can collapse their safety constraints and trigger unethical responses. In this work, we investigate jailbreaks i...

ROS-20251020-01

Vulnerability in open source external resource management software Terraform is associated with an incorrect restriction on the path name of a restricted directory. Exploitation of the vulnerability could allow an attacker to download arbitrary files...

JLSEC-2025-95

The SSH transport protocol with certain OpenSSH extensions, found in OpenSSH before 9.6 and other products, allows remote attackers to bypass integrity checks such that some packets are omitted from the extension negotiation message, and a client and server may consequently end up with a connecti...

OSV-2025-845 Global-buffer-overflow in icaltime_day_of_year

OSS-Fuzz report: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=452914686 Crash type: Global-buffer-overflow READ 4 Crash state: icaltimedayofyear expandyeardays iteratorsetstart...

Toeverything AFFiNE code injection vulnerability

Toeverything AFFiNE is a knowledge management software from Toeverything open source. AFFiNE 0.24.1 and earlier versions suffer from a code injection vulnerability that stems from an unknown code flaw in the Avatar Upload Image Endpoint component, which could lead to a cross-site scripting attack...

[SECURITY] Fedora 42 Update: firefox-144.0-3.fc42

Mozilla Firefox is an open-source web browser, designed for standards compliance, performance and portability...

SoK: Taxonomy and Evaluation of Prompt Security in Large Language Models

Large Language Models LLMs have rapidly become integral to real-world applications, powering services across diverse sectors. However, their widespread deployment has exposed critical security risks, particularly through jailbreak prompts that can bypass model alignment and induce harmful outputs...

DataEase cross-site scripting vulnerabilities

DataEase is an open source data visualization and analysis tool from DataEase Open Source. It is used to help users quickly analyze data and gain insight into business trends to achieve business improvement and optimization. A cross-site scripting vulnerability exists in DataEase 2.10.13 and...

ClipBucket path traversal vulnerability

ClipBucket is an open source and freely downloadable PHP script from MacWarrior Open Source. It is used for sharing video sites. A path traversal vulnerability exists in ClipBucket version 5.5.2 - 146, which stems from insufficient validation of file load paths and could lead to a path traversal...

DataEase code-related vulnerabilities

DataEase is a set of Java-based development of open source data visualization and analysis tools to help users quickly analyze data and insight into business trends , so as to achieve business improvement and optimization . A code injection vulnerability exists in DataEase DB2/MongoDB JDBC...

DataEase code-related vulnerabilities

DataEase is a set of Java-based development of open source data visualization and analysis tools to help users quickly analyze data and insight into business trends , so as to achieve business improvement and optimization . DataEase H2.java handles JDBC connection validation with a code injection...

Flowise security vulnerabilities

Flowise is a FlowiseAI open source tool for easily building LLM applications. A security vulnerability exists in Flowise 3.0.4 and earlier versions, which stems from a Supabase RPC Filter field that is not cleaned and escaped from user input, which could lead to remote code execution...

Dify cross-site scripting vulnerabilities

Dify is an open source LLM application development platform from LangGenius Open Source. A cross-site scripting vulnerability exists in Dify 1.9.1 and earlier versions, which stems from a failure to validate or clean up the authorizationurl in the implementation of the OAuth process, which could...

When Intelligence Fails: An Empirical Study on Why LLMs Struggle with Password Cracking

The remarkable capabilities of Large Language Models LLMs in natural language understanding and generation have sparked interest in their potential for cybersecurity applications, including password guessing. In this study, we conduct an empirical investigation into the efficacy of pre-trained LL...

CVE-2025-61909

Icinga 2 is an open source monitoring system. From 2.10.0 to before 2.15.1, 2.14.7, and 2.13.13, the safe-reload script also used during systemctl reload icinga2 and logrotate configuration shipped with Icinga 2 read the PID of the main Icinga 2 process from a PID file writable by the daemon user...

CVE-2025-61907

Icinga 2 is an open source monitoring system. In Icinga 2 versions 2.4 through 2.15.0, filter expressions provided to the various /v1/objects endpoints could access variables or objects that would otherwise be inaccessible for the user. This allows authenticated API users to learn information tha...

How to Get Started with NVIDIA cuOpt

A simple guide to get started with cuOpt, an open source, GPU-accelerated solver for decision optimization...

MAL-2025-48432 Malicious code in company-browser-package (npm)

--- -= Per source details. Do not edit below this line.=- Source: ossf-package-analysis 7621dd08044aeaacb68745078c793611d91031eb9852f8f667f739d485efe939 The OpenSSF Package Analysis project identified 'company-browser-package' @ 99.9.10 npm as malicious. It is considered malicious because: - The...

Leveraging Code Cohesion Analysis to Identify Source Code Supply Chain Attacks

Supply chain attacks significantly threaten software security with malicious code injections within legitimate projects. Such attacks are very rare but may have a devastating impact. Detecting spurious code injections using automated tools is further complicated as it often requires deciphering t...