ChatGPhish Vulnerability Turns ChatGPT Web Summaries Into a Phishing Surface

Cybersecurity researchers have disclosed details of a vulnerability in OpenAI ChatGPT that leverages the artificial intelligence AI assistant's implicit trust in Markdown links and images to trigger prompt injections and open the door to phishing attacks. The technique has been codenamed ChatGPhi...

Duplicate Advisory: OpenClaw: QQBot direct media upload skipped URL SSRF validation

Duplicate Advisory This advisory has been withdrawn because it is a duplicate of GHSA-c4qg-j8jg-42q5. This link is maintained to preserve external references. Original Description OpenClaw before 2026.4.20 contains a server-side request forgery vulnerability in QQBot direct media upload that skip...

CVE-2026-44117

OpenClaw before 2026.4.20 contains a server-side request forgery vulnerability in QQBot direct media upload that skips URL validation. Attackers can bypass SSRF protections by sending crafted image URLs to uploadC2CMedia and uploadGroupMedia endpoints to relay unintended requests...

CVE-2026-44117 OpenClaw < 2026.4.20 - Server-Side Request Forgery in QQBot Direct Media Upload

OpenClaw before 2026.4.20 contains a server-side request forgery vulnerability in QQBot direct media upload that skips URL validation. Attackers can bypass SSRF protections by sending crafted image URLs to uploadC2CMedia and uploadGroupMedia endpoints to relay unintended requests...

PT-2026-38250

Name of the Vulnerable Software and Affected Versions OpenClaw versions prior to 2026.4.20 Description An issue in QQBot direct media upload allows for server-side request forgery SSRF, a flaw where a server is tricked into making requests to an unintended location. This occurs because URL...

CVE-2026-2826

The Kadence Blocks — Page Builder Toolkit for Gutenberg Editor plugin for WordPress is vulnerable to authorization bypass in all versions up to, and including, 3.6.3. This is due to the plugin not properly verifying that a user has the uploadfiles capability in the processpattern REST API endpoin...

CVE-2026-2826 Kadence Blocks — Page Builder Toolkit for Gutenberg Editor <= 3.6.3 - Missing Authorization to Authenticated (Contributor+) Media Upload

The Kadence Blocks — Page Builder Toolkit for Gutenberg Editor plugin for WordPress is vulnerable to authorization bypass in all versions up to, and including, 3.6.3. This is due to the plugin not properly verifying that a user has the uploadfiles capability in the processpattern REST API endpoin...

CVE-2025-69727

An Incorrect Access Control vulnerability exists in INDEX-EDUCATION PRONOTE prior to 2025.2.8. The affected components index.js and composeUrlImgPhotoIndividu allow the construction of direct URLs to user profile images based solely on predictable identifiers such as user IDs and names. Due to...

CVE-2026-27023

Twenty is an open source CRM. Prior to version 1.18, the SSRF protection in SecureHttpClientService validated request URLs at the request level but did not validate redirect targets. An authenticated user who could control outbound request URLs e.g., webhook endpoints, image URLs could bypass...

Allocation of Resources Without Limits or Throttling

Overview jspdf is a PDF Document creation from JavaScript Affected versions of this package are vulnerable to Allocation of Resources Without Limits or Throttling via the addImage and html methods. An attacker can cause excessive memory allocation and application unavailability by supplying...

CVE-2026-26013

LangChain is a framework for building agents and LLM-powered applications. Prior to 1.2.11, the ChatOpenAI.getnumtokensfrommessages method fetches arbitrary imageurl values without validation when computing token counts for vision-enabled models. This allows attackers to trigger Server-Side Reque...

EUVD-2025-14023

Malicious code in bioql PyPI...

EUVD-2023-44532

Malicious code in bioql PyPI...

SUSE CVE-2025-50738

The Memos application, up to version v0.24.3, allows for the embedding of markdown images with arbitrary URLs. When a user views a memo containing such an image, their browser automatically fetches the image URL without explicit user consent or interaction beyond viewing the memo. This can be...

Cross-site Scripting (XSS)

github.com/usememos/memos is vulnerable to Cross-site Scripting XSS. The vulnerability is due to unrestricted external image embedding because markdown images with arbitrary URLs are automatically fetched when viewing a memo, exposing the user's IP address, browser User-Agent, and other...

Hugging Face Transformers 输入验证错误漏洞

Hugging Face Transformers is Hugging Face's open source advanced natural language processing for Jax, PyTorch and TensorFlow. An input validation error vulnerability exists in Hugging Face Transformers version 4.49.0 and earlier, which stems from insufficient URL validation in imageutils.py and...

CVE-2025-53018 Lychee has Server-Side Request Forgery (SSRF) in Photo::fromUrl API via unvalidated remote image URLs

Lychee is a free, open-source photo-management tool. Prior to version 6.6.13, a critical Server-Side Request Forgery SSRF vulnerability exists in the /api/v2/Photo::fromUrl endpoint. This flaw lets an attacker instruct the application’s backend to make HTTP requests to any URL they choose...

CVE-2023-29725

The BT21 x BTS Wallpaper app 12 for Android allows unauthorized applications to actively request permission to insert data into the database that records information about a user's personal preferences and will be loaded into memory to be read and used when the application is opened. By injecting...

CVE-2023-3906

An input validation issue in the asset proxy in GitLab EE, affecting all versions from 12.3 prior to 16.2.8, 16.3 prior to 16.3.5, and 16.4 prior to 16.4.1, allowed an authenticated attacker to craft image urls which bypass the asset proxy...

CVE-2025-29907

jsPDF is a library to generate PDFs in JavaScript. Prior to 3.0.1, user control of the first argument of the addImage method results in CPU utilization and denial of service. If given the possibility to pass unsanitised image urls to the addImage method, a user can provide a harmful data-url that...