Common Vulnerability Scoring System
PulseAugur coverage of Common Vulnerability Scoring System — every cluster mentioning Common Vulnerability Scoring System across labs, papers, and developer communities, ranked by signal.
9 day(s) with sentiment data
-
AI Safety Group Proposes Standardized Jailbreak Rating Framework
The Frontier Model Forum (FMF), a consortium including Anthropic, Google, Microsoft, and OpenAI, has introduced a new framework to evaluate AI jailbreaks. This framework, inspired by the Common Vulnerability Scoring Sys…
-
Multiple critical software vulnerabilities disclosed, impacting GitLab, WordPress, and more · 3 sources tracked
Multiple security vulnerabilities have been disclosed across various software platforms, with researchers highlighting their presence and potential impact. These include issues in SonicWall, OpenSSH, Barracuda ESG, and …
-
C3RB3R ransomware exploits Synology NAS vulnerability
The C3RB3R ransomware is actively targeting Synology Network Attached Storage (NAS) devices that are running outdated versions of their DiskStation Manager (DSM) software. This attack exploits a critical remote code exe…
-
Linux Kernel TCP Flaw Exploitable by AI, Critical Vulnerability Affects Major Distributions
A critical vulnerability (CVE-2026-98432) has been discovered in the Linux kernel's TCP implementation, with a CVSS score of 9.8, potentially allowing remote attackers to execute commands as root. Anthropic's AI model r…
-
Critical MCP Inspector vulnerabilities missed by npm audit
Critical security vulnerabilities have been identified in the MCP Inspector tool, specifically affecting versions up to 1.0.2 on the 1.x line and up to 2.12.0 on the 2.x line. These vulnerabilities, with CVSS scores of …
-
Google's Gemini 4 Carbon, AI agent trading issues, and security vulnerabilities highlighted · 4 sources tracked
Google has announced Gemini 4 Carbon, its most powerful model to date, on October 28, 2026. Separately, discussions around AI agents highlight potential issues with order placement on platforms like Alpaca, where agent-…
-
Critical PHP2Shell Vulnerabilities Demand Infrastructure-Level Fixes
Two critical vulnerabilities in PHP, collectively known as PHP2Shell, were disclosed on October 21, 2026, allowing unauthenticated attackers to execute code remotely. These flaws, with a CVSS 4.0 score of 9.2, reside wi…
-
Unofficial Qwen3.8 derivative shows benchmark gains, retains CVSS scoring issues
An unofficial derivative model, autotrust/Qwen3.8-Flash-Next-E384-FP4-V2, has been released, built upon Alibaba Group's Qwen3.8-Flash-Next-E384-FP4-V2. This model incorporates GRPO-style reinforcement learning and is va…
-
Critical @modelcontextprotocol/server SDK flaw exposes client data
A critical vulnerability, CVE-2026-111770, has been identified in the @modelcontextprotocol/server SDK, affecting versions 2.0.0 through 2.2.0. The flaw allows one client to receive another client's responses or notific…
-
Vibe-Trading AI agent vulnerable to root command execution via prompt injection
A critical prompt injection vulnerability, CVE-2026-114107, has been discovered in Vibe-Trading's AI trading-research agent, specifically in versions prior to 0.1.7. This vulnerability allows attackers to execute arbitr…
-
New framework uses digital twins to score cyber-social impact of attacks
Researchers have developed a new framework that combines digital twin technology with a Cyber-Social Security model to better analyze and quantify the impact of cyberattacks on critical infrastructures. This approach us…
-
Claude Code Desktop vulnerability allows data exfiltration via preview pane
Noma Labs has identified a critical vulnerability in Anthropic's Claude Code Desktop application, dubbed SneakPeek. The exploit allows malicious HTML files, disguised as code coverage reports and exceeding 256KB, to byp…
-
Cybersecurity faces 'negative time to exploit' as attackers outpace patching
The cybersecurity landscape has shifted dramatically, with exploitation of vulnerabilities now frequently occurring before a patch is even available, a trend highlighted by Mandiant's M-Trends 2026 report. This "negativ…
-
GitLab Patches Critical AI Gateway Vulnerability Allowing Sandbox Escape
GitLab has patched a critical vulnerability (CVSS 9.9) in its self-hosted AI gateway that allowed users to escape the prompt template sandbox. The flaw, identified as CVE-2026-90970, could enable malicious actors to exe…
-
MCP Python SDK OAuth flaw allows credential theft
A security vulnerability has been identified in the MCP Python SDK, affecting versions 1.9.1–1.29.1 and 2.0.0–2.1.1. The flaw allows a malicious MCP server to intercept an application's OAuth credentials, including clie…
-
CISA discontinues weekly vulnerability bulletin for risk-based approach · 2 sources tracked
The US Cybersecurity and Infrastructure Security Agency (CISA) has announced it will no longer publish its weekly vulnerability bulletin. This decision stems from a shift towards a more dynamic, risk-based prioritizatio…
-
Linux and open-source software receive critical security updates
This article discusses the recent updates and security patches for Linux and open-source software, highlighting an urgent hotfix with a critical CVSS score. It also touches upon the release of 113 packages and the chall…
-
New research tackles multilingual AI efficiency and capabilities
Researchers are developing new methods to improve the efficiency and capabilities of multilingual AI models. One study explores token merging for multilingual speech recognition, showing it can significantly reduce comp…
-
Critical CVSS-rated vulnerability prompts Debian and Void updates
A critical vulnerability, rated with a perfect CVSS score and no available workarounds, has been identified. Debian has released 107 updates, while Void has issued 113. The exact source of these updates remains unclear.
-
Critical Mastodon vulnerability (CVSS 10/10) allows server takeover via .prompty file
A critical vulnerability, rated 10 out of 10 on the Common Vulnerability Scoring System (CVSS), has been discovered in the Mastodon platform. This security flaw could be exploited through a seemingly harmless .prompty f…