Latest Vulnerabilities with Public Exploits

Incorrect access control in the /api/v1/conversations/*/files API of GT Edge AI Platform before v2.0.10 allows unauthorized attackers to access other users' uploaded files.

Insecure permissions in the /api/v1/agents API of GT Edge AI Platform before v2.0.10-dev allows unauthorized attackers to access sensitive information.

An issue in GT Edge AI Community Edition Versions before v2.0.12 allows attackers to execute arbitrary code via injecting a crafted JSON payload into the Prompt window.

LSC Smart Connect Indoor IP Camera 1.4.13 contains a RCE vulnerability in start_app.sh.

Tenda AC15 v15.03.05.18_multi) issues an authentication cookie that exposes the account password hash to the client and uses a short, low-entropy suffix as the session identifier. An attacker with network access or the ability to run JS in a victim browser can steal the cookie and replay it to access protected resources.

Claude Code is an agentic coding tool. Prior to version 1.0.39, when running on a machine with Yarn 3.0 or above, Claude Code could have been tricked to execute code contained in a project via yarn plugins before the user accepted the startup trust dialog. Exploiting this would have required a user to start Claude Code in an untrusted directory and to be using Yarn 3.0 or above. This issue has been patched in version 1.0.39.

Critical XXE in Apache Tika (tika-parser-pdf-module) in Apache Tika 1.13 through and including 3.2.1 on all platforms allows an attacker to carry out XML External Entity injection via a crafted XFA file inside of a PDF. An attacker may be able to read sensitive data or trigger malicious requests to internal resources or third-party servers. Note that the tika-parser-pdf-module is used as a dependency in several Tika packages including at least: tika-parsers-standard-modules, tika-parsers-standard-package, tika-app, tika-grpc and tika-server-standard. Users are recommended to upgrade to version 3.2.2, which fixes this issue.

The JAY Login & Register plugin for WordPress is vulnerable to authentication bypass in versions up to, and including, 2.4.01. This is due to incorrect authentication checking in the 'jay_login_register_process_switch_back' function with the 'jay_login_register_process_switch_back' cookie value. This makes it possible for unauthenticated attackers to log in as any existing user on the site, such as an administrator, if they have access to the user id.

The Fox LMS – WordPress LMS Plugin plugin for WordPress is vulnerable to privilege escalation in all versions up to, and including, 1.0.5.1. This is due to the plugin not properly validating the 'role' parameter when creating new users via the `/fox-lms/v1/payments/create-order` REST API endpoint. This makes it possible for unauthenticated attackers to create new user accounts with arbitrary roles, including administrator, leading to complete site compromise.

1C-Bitrix through 25.100.500 allows Remote Code Execution because an actor with SOURCE/WRITE permissions for the Translate Module can upload and execute code by sending a PHP file and a .htaccess file. NOTE: this is disputed by the Supplier because this is intended behavior for the high-privileged users who can upload new translated pages to the website.

A local privilege escalation vulnerability due to insufficient authorization in the SonicWall SMA1000 appliance management console (AMC).

Bitrix24 through 25.100.300 allows Remote Code Execution because an actor with SOURCE/WRITE permissions for the Translate Module can upload and execute code by sending a PHP file and a .htaccess file. NOTE: this is disputed by the Supplier because this is intended behavior for the high-privileged users who can upload new translated pages to the website.

Open Source Point of Sale (opensourcepos) is a web based point of sale application written in PHP using CodeIgniter framework. Starting in version 3.4.0 and prior to version 3.4.2, a Cross-Site Request Forgery (CSRF) vulnerability exists in the application's filter configuration. The CSRF protection mechanism was **explicitly disabled**, allowing the application to process state-changing requests (POST) without verifying a valid CSRF token. An unauthenticated remote attacker can exploit this by hosting a malicious web page. If a logged-in administrator visits this page, their browser is forced to send unauthorized requests to the application. A successful exploit allows the attacker to silently create a new Administrator account with full privileges, leading to a complete takeover of the system and loss of confidentiality, integrity, and availability. The vulnerability has been patched in version 3.4.2. The fix re-enables the CSRF filter in `app/Config/Filters.php` and resolves associated AJAX race conditions by adjusting token regeneration settings. As a workaround, administrators can manually re-enable the CSRF filter in `app/Config/Filters.php` by uncommenting the protection line. However, this is not recommended without applying the full patch, as it may cause functionality breakage in the Sales module due to token synchronization issues.

Heap-based buffer overflow in Windows Common Log File System Driver allows an authorized attacker to elevate privileges locally.

Heap-based buffer overflow in Windows Cloud Files Mini Filter Driver allows an authorized attacker to elevate privileges locally.

Chatwoot is a customer engagement suite. Prior to 3.16.0, conversation and contact filters endpoints did not sanitize the input of query_operator passed from the frontend or the API. This provided any actor who is authenticated, an attack vector to run arbitrary SQL within the filter query by adding a tautological WHERE clause. This issue is patched with v3.16.0.

Cybersecurity AI (CAI) is an open-source framework for building and deploying AI-powered offensive and defensive automation. Versions 0.5.9 and below are vulnerable to Command Injection through the run_ssh_command_with_credentials() function, which is available to AI agents. Only password and command inputs are escaped in run_ssh_command_with_credentials to prevent shell injection; while username, host and port values are injectable. This issue does not have a fix at the time of publication.

FileRise is a self-hosted web file manager / WebDAV server. Versions prior to 2.7.1 are vulnerable to Stored Cross-Site Scripting (XSS) due to unsafe handling of browser-renderable user uploads when served through the sharing and download endpoints. An attacker who can get a crafted SVG (primary) or HTML (secondary) file stored in a FileRise instance can cause JavaScript execution when a victim opens a generated share link (and in some cases via the direct download endpoint). This impacts share links (`/api/file/share.php`) and direct file access / download path (`/api/file/download.php`), depending on browser/content-type behavior. Version 2.7.1 fixes the issue.

OrangeHRM is a comprehensive human resource management (HRM) system. From version 5.0 to 5.7, the application contains an input-neutralization flaw in its mail configuration and delivery workflow that allows user-controlled values to flow directly into the system’s sendmail command. Because these values are not sanitized or constrained before being incorporated into the command execution path, certain sendmail behaviors can be unintentionally invoked during email processing. This makes it possible for the application to write files on the server as part of the mail-handling routine, and in deployments where those files end up in web-accessible locations, the behavior can be leveraged to achieve execution of attacker-controlled content. The issue stems entirely from constructing OS-level command strings using unsanitized input within the mail-sending logic. This issue has been patched in version 5.8.

NVIDIA CUDA Toolkit for all platforms contains a vulnerability in cuobjdump where an attacker may cause a stack-based buffer overflow by getting the user to run cuobjdump on a malicious ELF file. A successful exploit of this vulnerability may lead to arbitrary code execution at the privilege level of the user running cuobjdump.