SUSE CVE-2026-7907

Use after free in DOM in Google Chrome prior to 148.0.7778.96 allowed a remote attacker to execute arbitrary code inside a sandbox via a crafted HTML page. Chromium security severity: High...

Hard to Read, Easy to Jailbreak: How Visual Degradation Bypasses MLLM Safety Alignment

Recent advancements in visual context compression enable MLLMs to process ultra-long contexts efficiently by rendering text into images. However, we identify a critical vulnerability inherent to this paradigm: lowering image resolution inadvertently catalyzes jailbreaking. Our experiments reveal...

CVE-2023-42345

A Cross Site Scripting vulnerability in Alkacon OpenCms before 16 exists via updateModelGroups.jsp...

Can I Check What I Designed? Mapping Security Design DSLs to Code Analyzers

When assessing the potential impact of code-level vulnerabilities, e.g., discovered by automated analyzers, it is essential to consider them in the context of the system's security design. However, this is a challenging task due to the abstraction gap between security design, often specified usin...

n8n-MCP 安全漏洞

n8n-MCP is a model context protocol server developed by Romuald Członkowski, an individual developer. It serves as a connection between AI assistants and automated workflow platforms. Versions of n8n-MCP from 2.18.7 to 2.50.2 contained security vulnerabilities. These vulnerabilities were caused b...

PT-2026-39294

Name of the Vulnerable Software and Affected Versions Open WebUI versions 0.3.5 through 0.8.12 Description A stored cross-site scripting XSS issue allows authenticated users with model creation permissions workspace.models to execute arbitrary JavaScript in the browsers of other users, including...

PT-2026-39272

Name of the Vulnerable Software and Affected Versions Open WebUI versions prior to 0.9.0 Description Open WebUI supports model composition through the base model id variable, allowing a user-defined model to reference a base model for inference. An access control flaw exists where the system...

PT-2026-39273

Name of the Vulnerable Software and Affected Versions Open WebUI versions prior to 0.9.0 Description The '/responses' endpoint in the OpenAI router allows any authenticated user to forward requests to upstream LLM providers without enforcing per-model access control. While the generate chat...

PT-2026-39083

Name of the Vulnerable Software and Affected Versions Linux kernel affected versions not specified Description A NULL pointer dereference occurs in the legacy NCM driver within the gncm bind function. This issue arises because the driver attempts to access the net device before it is fully...

SecureForge: Finding and Preventing Vulnerabilities in LLM-Generated Code Via Prompt Optimization

LLM coding agents now generate code at an unprecedented scale, yet LLM-generated code introduces cybersecurity vulnerabilities into codebases without human involvement. Even when frontier models are explicitly asked to write secure production code with relevant weaknesses to avoid in context, we...

PT-2026-39280

Name of the Vulnerable Software and Affected Versions Open WebUI versions prior to 0.9.0 Description Four Ollama proxy endpoints accept any model name from the user and forward the request to the Ollama backend without verifying if the user is authorized to access that model. While these endpoint...

Prompt Injection

Overview org.springframework.ai:spring-ai-model is a Core model interfaces and classes for Spring AI Affected versions of this package are vulnerable to Prompt Injection via conversation memory handling in the affected advisor. An attacker can inject crafted input in conversation memory that is...

n8n-MCP 日志信息泄露漏洞

n8n-MCP is a model context protocol server developed by Romuald Członkowski, an individual developer. Versions of n8n-MCP prior to 2.47.13 contained a vulnerability related to log information leakage. This vulnerability occurred because the complete parameters of MCP tool calls and JSON-RPC...

PT-2026-39003

Name of the Vulnerable Software and Affected Versions PraisonAI versions prior to 4.6.34 Description The Model Context Protocol MCP server in PraisonAI contains a path traversal flaw in its file-handling tools. The server registers four tools by default: 'praisonai.rules.create',...

Timing Attack

Overview mcp-ssh-tool is a Model Context Protocol MCP SSH client server for remote automation Affected versions of this package are vulnerable to Timing Attack in the transfer-related filesystem handling process. An attacker can access unauthorized files or directories by bypassing local path...

Exposure of Private Personal Information to an Unauthorized Actor

Overview Affected versions of this package are vulnerable to Exposure of Private Personal Information to an Unauthorized Actor via the Email field in the Comment model exposed through unauthenticated public API endpoints. An attacker can obtain the email addresses of all guest commenters by makin...

GHSA-RJ4G-RQGH-RX9H Ech0 comment model's Email field returned on public /api/comments endpoints

Summary The Comment model serializes its Email field through the public comment-listing API. internal/model/comment/comment.go:33 uses json:"email", while adjacent PII fields IPHash, UserAgent correctly use json:"-". The public endpoints GET /api/comments?echoid=X and GET...

Ech0 comment model's Email field returned on public /api/comments endpoints

Summary The Comment model serializes its Email field through the public comment-listing API. internal/model/comment/comment.go:33 uses json:"email", while adjacent PII fields IPHash, UserAgent correctly use json:"-". The public endpoints GET /api/comments?echoid=X and GET...

Exposure of Private Personal Information to an Unauthorized Actor

Overview Affected versions of this package are vulnerable to Exposure of Private Personal Information to an Unauthorized Actor via the Email field in the Comment model exposed through unauthenticated public API endpoints. An attacker can obtain the email addresses of all guest commenters by makin...

How to Disable Google's Gemini in Chrome

Chrome users were caught off guard by a 4-GB Google AI model baked into Chrome, sparking privacy concerns. The good news: You can easily uninstall it. The bad? You might not want to...