| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| Malicious content uploaded as file can be used to execute script code when following attacker-controlled links. Unintended actions can be executed in the context of the users account, including exfiltration of sensitive information. Please deploy the provided updates and patch releases. No publicly available exploits are known |
| llama-cpp-python is the Python bindings for llama.cpp. `llama-cpp-python` depends on class `Llama` in `llama.py` to load `.gguf` llama.cpp or Latency Machine Learning Models. The `__init__` constructor built in the `Llama` takes several parameters to configure the loading and running of the model. Other than `NUMA, LoRa settings`, `loading tokenizers,` and `hardware settings`, `__init__` also loads the `chat template` from targeted `.gguf` 's Metadata and furtherly parses it to `llama_chat_format.Jinja2ChatFormatter.to_chat_handler()` to construct the `self.chat_handler` for this model. Nevertheless, `Jinja2ChatFormatter` parse the `chat template` within the Metadate with sandbox-less `jinja2.Environment`, which is furthermore rendered in `__call__` to construct the `prompt` of interaction. This allows `jinja2` Server Side Template Injection which leads to remote code execution by a carefully constructed payload. |
| XMLUnit for Java before 2.10.0, in the default configuration, might allow code execution via an untrusted stylesheet (used for an XSLT transformation), because XSLT extension functions are enabled. |
| Paymenter is a free and open-source webshop solution for hostings. Prior to version 1.2.11, the ticket attachments functionality in Paymenter allows a malicious authenticated user to upload arbitrary files. This could result in sensitive data extraction from the database, credentials being read from configuration files, and arbitrary system commands being run under the web server user context. This vulnerability was patched by commit 87c3db4 and was released under the version 1.2.11 tag without any other code modifications compared to version 1.2.10. If upgrading is not immediately possible, administrators can mitigate this vulnerability with one or more of the following measures: updating nginx config to download attachments instead of executing them or disallowing access to /storage/ fully using a WAF such as Cloudflare. |
| The Runtime components of messageformat package for Node.js before 3.0.2 contain a prototype pollution vulnerability. Due to insufficient validation of nested message keys during the processing of message data, an attacker can manipulate the prototype chain of JavaScript objects by providing specially crafted input. This can result in the injection of arbitrary properties into the Object.prototype, potentially leading to denial of service conditions or unexpected application behavior. The vulnerability allows attackers to alter the prototype of base objects, impacting all subsequent object instances throughout the application's lifecycle. |
| LinkJoin through 882f196 mishandles token ownership in password reset. |
| ERC (aka Emotion Recognition in Conversation) through 0.3 has insecure deserialization via a serialized object because jsonpickle is used. |
| In Agora Foundation Agora fall23-Alpha1 before 690ce56, there is XSS via a profile picture to server/controller/userController.js. Formats other than PNG, JPEG, and WEBP are permitted by server/routes/userRoutes.js; this includes SVG. |
| In Agora Foundation Agora fall23-Alpha1 before b087490, there is XSS via tag in client/agora/public/js/editorManager.js. |
| An integer overflow vulnerability in the loading of ExecuTorch models can cause smaller-than-expected memory regions to be allocated, potentially resulting in code execution or other undesirable effects. This issue affects ExecuTorch prior to commit 8f062d3f661e20bb19b24b767b9a9a46e8359f2b. |
| The web application is not protected against cross-site request forgery attacks. Therefore, an attacker can trick users into performing actions on the application when they visit an attacker-controlled website or click on a malicious link. E.g. an attacker can forge malicious links to reset the admin password or create new users. |
| A group of related buffer overflow vulnerabilities in the loading of ExecuTorch models can cause the runtime to crash and potentially result in code execution or other undesirable effects. This issue affects ExecuTorch prior to commit cea9b23aa8ff78aff92829a466da97461cc7930c. |
| The scanner device boots into a kiosk mode by default and opens the Scan2Net interface in a browser window. This browser is run with the permissions of the root user. There are also several other applications running as root user. This can be confirmed by running "ps aux" as the root user and observing the output. |
| An out-of-bounds access vulnerability in the loading of ExecuTorch models can cause the runtime to crash and potentially result in code execution or other undesirable effects. This issue affects ExecuTorch prior to commit b6b7a16df5e7852d976d8c34c8a7e9a1b6f7d005. |
| The www-data user can elevate its privileges because sudo is configured to allow the execution of the mount command as root without a password. Therefore, the privileges can be escalated to the root user. The risk has been accepted by the vendor and won't be fixed in the near future. |
| A heap buffer overflow vulnerability in the loading of ExecuTorch models can potentially result in code execution or other undesirable effects. This issue affects ExecuTorch prior to commit ede82493dae6d2d43f8c424e7be4721abe5242be |
| Thinbus Javascript Secure Remote Password is a browser SRP6a implementation for zero-knowledge password authentication. In versions 2.0.0 and below, a protocol compliance bug causes the client to generate a fixed 252 bits of entropy instead of the intended bit length of the safe prime (defaulted to 2048 bits). The client public value is being generated from a private value that is 4 bits below the specification. This reduces the protocol's designed security margin it is now practically exploitable. The servers full sized 2048 bit random number is used to create the shared session key and password proof. This is fixed in version 2.0.1. |
| Let's Encrypt client and ACME library written in Go (Lego). In versions 4.25.1 and below, the github.com/go-acme/lego/v4/acme/api package (thus the lego library and the lego cli as well) don't enforce HTTPS when talking to CAs as an ACME client. Unlike the http-01 challenge which solves an ACME challenge over unencrypted HTTP, the ACME protocol requires HTTPS when a client communicates with the CA to performs ACME functions. However, the library fails to enforce HTTPS both in the original discover URL (configured by the library user) and in the subsequent addresses returned by the CAs in the directory and order objects. If users input HTTP URLs or CAs misconfigure endpoints, protocol operations occur over HTTP instead of HTTPS. This compromises privacy by exposing request/response details like account and request identifiers to network attackers. This was fixed in version 4.25.2. |
| A vulnerability was found in WAVLINK QUANTUM D2G, QUANTUM D3G, WL-WN530G3A, WL-WN530HG3, WL-WN532A3 and WL-WN576K1 up to V1410_240222 and classified as critical. Affected by this issue is the function sys_login of the file /cgi-bin/login.cgi of the component HTTP POST Request Handler. The manipulation of the argument login_page leads to buffer overflow. The attack may be launched remotely. The exploit has been disclosed to the public and may be used. The vendor was contacted early about this disclosure but did not respond in any way. |
| NiceGUI is a Python-based UI framework. Versions 2.24.2 and below are at risk for Cross-Site Scripting (XSS) when developers render unescaped user input into the DOM using ui.html(). NiceGUI did not enforce HTML or JavaScript sanitization, so applications that directly combine components like ui.input() with ui.html() or ui.chat_message with HTML content without escaping may allow attackers to execute arbitrary JavaScript in the user’s browser. Applications that do not pass untrusted input into ui.html() are not affected. This issue is fixed in version 3.0.0. |
