The master class on painting Christmas tree decorations was a success. 👍 They said that you can draw anything. 😏 I drew Elevation of Privilege (Local Privilege Escalation) – Linux Kernel (CVE-2024-1086). 😅
Qualys announced the TotalAI module for artificial intelligence (AI) and large language models (LLM) security. The module will be available in Q4 2024 as part of the Enterprise TruRisk platform.
Announced features:
🔹 Detection and monitoring of the AI infrastructure of organizations. To avoid “shadow LLM”.
🔹 Vulnerability Management with a focus on AI threats. Especially on countering theft (extraction) of data and models. They will offer a variety of ways to fix vulnerabilities.
🔹 Specialized LLM scanning focussed on prompt injection, model theft, and disclosure of confidential information.
🔹 Compliance Management and risk management. They emphasize combating data leaks and mention GDPR, PCI, CCPA.
There is a screenshot of the interface with statistics on models and related threats. We can also see statistics on threats related to assets and interesting informers for AI Workloads, AI Software and GPU.
Malware delivery via automatic software updates. An interesting case was published by Volexity experts. Let’s say you have some software on the host. It periodically goes online to check for updates and then installs them… And instead of updating, malware is installed. 😱👾
And this is NOT a case of a compromised software vendor website. 🙂
To achieve this result, the StormBamboo bad guys carried out a DNS poisoning attack at the Internet service provider (ISP) level. The attackers changed the responses to DNS queries for certain domains. Accordingly, the “update” was downloaded from the IP addresses of the villains. 😈
Naturally, for this trick to work, the software must use insecure update mechanisms: connect via HTTP, do not check installer digital signatures, etc. 🤷♂️ Volexity mentions an attack via 5KPlayer media player updates. The victims were infected with MACMA and POCOSTICK / MGBot malware.
I also made a meme with the cool Yusuf Dikeç. 😅
🔹 Every vulnerability existing in the infrastructure must be detected.
🔹 For each detected vulnerability, a patching task must be created.
This is the base. And when they tell you that you don’t have to do this because there is some super-modern vulnerability assessment and prioritization tool, you should be skeptical. 😉
Regarding the Qualys Patch Management event that took place yesterday.
I liked:
✅ Cool report by Eran Livne about Patch Management capabilities in Qualys. 👍 Especially about creating linked patching tasks (first for a test scope, and a week later for a full scope) and about the ability to isolate hosts as a mitigation option (access remains only from the Qualys cloud). The part about new TruRisk Eliminate was also interesting.
✅ Adam Gray beautifully justified the need for mandatory patching (since prevention doesn’t really work 🤷♂️).
I didn’t like:
❌ Most speakers focused on other information security topics rather than patch management. I think it would have been possible to select more thematic reports for this event.
❌ I simply can’t accept theses like “you don’t need to patch all vulnerabilities”. 🤷♂️ My position: you need to patch everything. And workarounds are good for a while UNTIL a patch is installed.
Remote Code Execution – Acronis Cyber Infrastructure (CVE-2023-45249). Due to the default passwords used, a remote unauthenticated attacker can gain access to an Acronis Cyber Infrastructure (ACI) server and execute arbitrary code.
ACI is a hyperconverged platform for storage, backup, computing, virtualization and networking.
🔻 Patches that fix this vulnerability were released on October 30, 2023 (❗️).
🔻 After 9-10 months, on July 24 of this year, Acronis noted in a bulletin that the vulnerability was exploited in the wild. The purpose of exploitation was to install a cryptominer. On July 29, the vulnerability was added to the CISA KEV.
Some sources report 20,000 service providers using ACI. I have not found any confirmation of this. Perhaps there is confusion with Acronis Cyber Protect. However, there are probably quite a few large companies using ACI. If you work for such a company, be sure to pay attention.
Remote Code Execution vulnerability – Artifex Ghostscript (CVE-2024-29510). Memory corruption allows to bypass the SAFER sandbox and execute arbitrary code.
Ghostscript is a PostScript and PDF document interpreter. It is used in ImageMagick, LibreOffice, GIMP, Inkscape, Scribus, CUPS, etc. It is available for many OS.
🔻 Ghostscript version 10.03.1, which fixes the vulnerability, was released on May 2.
🔻 On July 2, Codean Labs published a detailed analysis of this vulnerability and PoC. In the video they launch the calculator by opening a special ps file with the ghostscript utility or a special odt file in LibreOffice.
🔻 On July 10, a functional exploit was released on GitHub. And on July 19, a module for Metasploit was released.
👾 The media writes that the vulnerability is being exploited in the wild. However, it’s based on a single microblog post by some Portland developer. 🤷♂️ I think more reliable evidence of exploitation in attacks will appear soon.
This is my personal blog. The opinions expressed here are my own and not of my employer. All product names, logos, and brands are property of their respective owners. All company, product and service names used here for identification purposes only. Use of these names, logos, and brands does not imply endorsement. You can freely use materials of this site, but it would be nice if you place a link on https://avleonov.com and send message about it at me@avleonov.com or contact me any other way.
