Tag
By Waqas Millions of IoT and industrial devices at risk! Critical vulnerabilities in Cinterion cellular modems allow remote attackers to take control. This is a post from HackRead.com Read the original post: Cinterion Modem Vulnerabilities Leave IoT and Industrial Networks Exposed
Debian Linux Security Advisory 5687-1 - A security issue was discovered in Chromium, which could result in the execution of arbitrary code, denial of service or information disclosure. Google is aware that an exploit for CVE-2024-4671 exists in the wild.
By Deeba Ahmed Encrypted email services like ProtonMail and Wire promise privacy, but can they guarantee anonymity? A recent case in Spain has users questioning the limitations of encryption when law enforcement steps in. This is a post from HackRead.com Read the original post: Police Accessed Proton Mail User Data in Terrorism Probe
## Description `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 RCE by a carefully constructed payload. ## Source-to-Sink ### `llama.py` -> `class Llama` -> `__init__`: ```python class Llama: """High-level Python wrapper for a ...
The financially motivated threat actor known as FIN7 has been observed leveraging malicious Google ads spoofing legitimate brands as a means to deliver MSIX installers that culminate in the deployment of NetSupport RAT. "The threat actors used malicious websites to impersonate well-known brands, including AnyDesk, WinSCP, BlackRock, Asana, Concur, The Wall
By Deeba Ahmed Researchers uncover a novel cyberattack scheme called "LLMjacking" exploiting stolen cloud credentials to hijack powerful AI models. This article explores the implications of attackers leveraging large language models (LLMs) for malicious purposes and offers security recommendations for the cloud and AI communities. This is a post from HackRead.com Read the original post: New LLMjacking Attack Lets Hackers Hijack AI Models for Profit
### Summary The latest version of lobe-chat(by now v0.141.2) has an unauthorized ssrf vulnerability. An attacker can construct malicious requests to cause SSRF without logging in, attack intranet services, and leak sensitive information. ### Details * visit https://chat-preview.lobehub.com/settings/agent * you can attack all internal services by /api/proxy and get the echo in http response :)    ### PoC ```http POST /api/proxy HTTP/2 Host: xxxxxxxxxxxxxxxxx Cookie: LOBE_LOCALE=zh-CN; LOBE_THEME_PRIMARY_COLOR=undefined; LOBE_THEME_NEUTRAL_COLOR=undefined; _ga=GA1.1.86608329.1711346216; _ga_63LP1TV70T=GS1.1.1711346215.1.1.1711346846.0.0.0 Content-Length: 23 Sec-Ch-Ua: "Google Chrome";v=...
Google on Thursday released security updates to address a zero-day flaw in Chrome that it said has been actively exploited in the wild. Tracked as CVE-2024-4671, the high-severity vulnerability has been described as a case of use-after-free in the Visuals component. It was reported by an anonymous researcher on May 7, 2024. Use-after-free bugs, which arise when a program
Malicious Android apps masquerading as Google, Instagram, Snapchat, WhatsApp, and X (formerly Twitter) have been observed to steal users' credentials from compromised devices. "This malware uses famous Android app icons to mislead users and trick victims into installing the malicious app on their devices," the SonicWall Capture Labs threat research team said in a recent report. The
Ubuntu Security Notice 6766-1 - It was discovered that the Open vSwitch implementation in the Linux kernel could overflow its stack during recursive action operations under certain conditions. A local attacker could use this to cause a denial of service. Sander Wiebing, Alvise de Faveri Tron, Herbert Bos, and Cristiano Giuffrida discovered that the Linux kernel mitigations for the initial Branch History Injection vulnerability were insufficient for Intel processors. A local attacker could potentially use this to expose sensitive information.