Hello and welcome to the Friday, November 13, 2026 edition of the SANS Internet Storm Center's Stormcast. My name is Johannes Ullrich, recording today from Jacksonville, Florida. This episode is brought to you by the SANS.edu Bachelor's Degree Program in Applied Cybersecurity. Well and in diaries today I wrote up some attacks that we're seeing against GPUStack. GPUStack is a system that allows you to essentially make available clusters of GPU for AI workloads. As often these proxies that are in front of some of these systems have their issues and that's again sort of what hits us here. Now in this particular case we do have a special header that's being added by the proxy X-GPUStack-Real-IP. Many other proxies have similar headers and they always come with the big caveat, well it's possible for an attacker to spoof those headers. In this particular case if an attacker spoofs that the real IP is the loopback IP address, well in that case the attacker is being given access to the system and allowed to bypass authentication. So very straightforward to exploit vulnerability. Makes you wonder why it took them so long to really scan for it on a larger scale. No idea how many of these systems are actually exposed. Now a little bit tricky to sort of date this vulnerability. Apparently the fixed version of GPUStack was released back in July. The actual CVE number appears to have been assigned about a week ago. So I guess anywhere sort of in that time frame. There was also a GitHub page that had details about this vulnerability that's also sort of a couple months old. Again exploitation is trivial. No real sort of exploit really required. Just that specific header and of course an exposed system. And ConnectWise released another ScreenConnect security advisory. Now this one is less about a specific vulnerability in ScreenConnect, but more about an ongoing security investigation that ConnectWise is running given the compromise of some of their customer and partner sites. So they're looking currently at 400 compromised environments. They think they have it narrowed down to a number of third party plugins that may be responsible for some of these issues. And they disabled them now. Not really clear what the user is supposed to do here other than making sure that you do have the latest ScreenConnect client. That's version 26.6.5. That particular version does also fix a vulnerability in ScreenConnect itself that is still being employed by these compromises in order to then further spread malware and compromise additional systems in your network. Remember ScreenConnect is often used by support functions to actually reach out to remote systems. And that's essentially sort of what they are abusing here. Definitely make sure that you're up to date and also take a look at this latest advisory from ConnectWise as it has additional details as to what to look for in your particular environment as well as the list of the third party extensions. And that have now been disabled. And while I already talked twice in the last week about N-able and about the patches that were released now and that are available to patch vulnerabilities that sort of had been made public without first notifying N-able. We now have exploits publicly available for at least three of these vulnerabilities. There is a GitHub repository that has the code for these. I haven't verified them yet. So as always treat them with caution. Never really know what they are doing and haven't really looked at them close enough to tell you that they're safe to the person running them. But if you're running N-able definitely you must upgrade at this point. It's kind of at the point where you probably want to assume compromise if you have an exposed unpatched N-able system connected. And Microsoft today had a significant availability event where good parts of Microsoft 365 were not reachable. This did affect Exchange, Outlook, lots and lots of different parts of Microsoft 365 that were not available. As I'm recording this, they say that they're bringing systems back online, but it's still unavailable for a lot of users. This also spilled over to many other sites. For example, if you relied on logging in via Microsoft 365's single sign-on system, Entra and such. This apparently was also affected by this outage. Even the Microsoft status page I heard was at times unavailable. Maybe just because of the high load being directed at that particular site, even though that's actually a cloud hosted site. So not sure if this is part of the outage or part of the high load. So if you're running Microsoft 365 Business or Enterprise, that's why your email wasn't working today. And hopefully by the time you're listening to this podcast, things will look better. Well, and that's it for today. Thanks for listening. Thanks for liking. Thanks for recommending. And thanks for subscribing to this podcast and talk to you again on Monday. Bye. Bye.