Podcast Detail

NGINX2ROOT HTTP/2 RCEs (CVE-2026-123007, CVE-2026-123633); Microsoft Agent Framework PyPI Compromise; Holiday Hack Challenge (Nov 19 2026)

If you are not able to play the podcast using the player below: Use this direct link to the audio file: https://traffic.libsyn.com/securitypodcast/10186.mp3

Podcast Logo
NGINX2ROOT HTTP/2 RCEs (CVE-2026-123007, CVE-2026-123633); Microsoft Agent Framework PyPI Compromise; Holiday Hack Challenge (Nov 19 2026)
00:00

My Next Class

Click HERE to learn more about classes Johannes is teaching for SANS

Today's episode covers F5's NGINX2ROOT HTTP/2 remote code execution patches, a malicious release of Microsoft's Agent Framework on the Python Package Index (PyPI), and the launch of the SANS Holiday Hack Challenge.

NGINX2ROOT HTTP/2 Remote Code Execution (CVE-2026-123007, CVE-2026-123633)
F5 patched five NGINX vulnerabilities, including critical HTTP/2 memory-corruption bugs that can allow unauthenticated remote code execution. The most widely exposed issue, CVE-2026-123007, does not require a special NGINX configuration beyond HTTP/2 being enabled, though exploitation needs a large volume of traffic and is not always reliable, even when Address Space Layout Randomization (ASLR) can be bypassed. Researchers showed these bugs can be chained with a kernel privilege-escalation vulnerability to obtain root access. Rapid7 released a proof-of-concept that works against NGINX running in containers; until you can patch, consider turning off HTTP/2 (Hypertext Transfer Protocol version 2).
https://my.f5.com/manage/s/article/K000163926
https://www.wiz.io/blog/nginx-http2-rce-vulnerabilities
https://github.com/sfewer-r7/CVE-2026-123007

Microsoft Agent Framework Compromised on PyPI
Wiz reported that malicious versions 1.26.1 and 1.26.2 of Microsoft's Agent Framework were published to the Python Package Index (PyPI) on November 17 after a compromised publishing credential was used from a Microsoft IP address. Both malicious versions have been removed; 1.26.0 is the latest safe release. The package normally sees about 41,000 downloads a week, and the malicious builds were only available for part of that day.
https://www.wiz.io/blog/malicious-version-of-microsofts-agent-framework-published-to-pypi

My Upcoming Classes
https://www.sans.org/profiles/dr-johannes-ullrich

SANS Holiday Hack Challenge
The SANS Holiday Hack Challenge is live again, with links in the show notes and across the SANS website. It may be a welcome distraction between patching and incident work.
https://www.sans.org/cyber-ranges/holiday-hack-challenge

Podcast Transcript

 Hello and welcome to the Thursday, November 19th, 2026
 edition of the SANS Internet Storm Center's Stormcast. My
 name is Johannes Ullrich, recording today from
 Jacksonville, Florida. And this episode is brought to you
 by the SANS.edu Bachelor's Degree Program in Applied
 Cybersecurity. Well, not much to talk about today regarding
 diaries. I expanded a little bit on the remarks I made at
 the end of yesterday's podcast and explained why we aren't
 going to yellow again, even though vulnerabilities like
 Java4Shell or NGINX2Shell, which we'll talk about
 next, are certainly sort of the same caliber as the PHP2Shell
 vulnerability a month ago. But it's really just
 about the new normal and higher Infocon is supposed to
 indicate if something that's not normal that requires
 additional attention is happening. And well, that's as
 far as the current state of the Internet and security
 goes, not the case right now. Well, so let's talk about
 NGINX. It's actually NGINX2ROOT, not to shell, as I
 mentioned earlier. The reason they call it NGINX2ROOT is
 that Wiz in their write-up actually demonstrated how you
 can take this all the way even with address space layout
 randomization by then also in the end leveraging kernel
 privilege escalation vulnerability all the way to
 obtain root access on the system via NGINX. It does not
 require any specific configuration for NGINX, but
 it does require quite a bit of traffic, about 256 gigabytes
 is sort of what's being reported from some of the
 proof-of-concept exploits that have been released. And yes,
 Rapid7 has released a proof-of-concept. As many of these
 memory management vulnerabilities, exploitation
 is a little bit iffy, so it's not sort of a super reliable
 exploit even though address space layout randomization
 can be bypassed using these vulnerabilities. Total of five
 vulnerabilities being patched here. Not all of them affect
 the default configuration. Some affect things like
 proxies and the like set up with NGINX. Now there are two
 organizations credited with discovering the vulnerability,
 Wiz and Xint. The real discovery here is modern AI
 models and that's what we have these days that we often have
 these multiple organizations finding these vulnerabilities
 at the same time, which of course means there may be
 other organizations that didn't report it to a vendor
 that may also have found these vulnerabilities and have
 working exploits. The working Rapid7 proof-of-concept
 exploit only works against NGINX running in containers.
 So again, a little bit more constrained environment here
 where some of these exploits tend to work better, tend to
 be more reliable than sort of your average NGINX install if
 there is one like that. Well, a quick answer: patch. And I don't
 like to recommend it, but maybe turn off HTTP/2. I'll leave
 that up to you. That's something that researchers
 really have sort of focused in on lately. We also had some
 Apache vulnerabilities with HTTP/2 and it is a complex
 protocol. So in particular from a memory management's
 point of view, there are probably a couple of
 undiscovered vulnerabilities out there. And then Wiz is
 also reporting that Microsoft's Agent Framework
 was compromised on PyPI. This apparently happened yesterday
 and the root cause, according to Wiz, was a compromised
 deployment key that was leaked. Now, interestingly,
 initially the publication of the malicious version actually
 failed because it came from the wrong IP address. But
 after using a Microsoft IP address, they were able to
 publish the malicious version of this Agent Framework. Not
 really clear how much of an effort it took to get the
 right IP address, given that Microsoft, with their cloud
 properties and such, if they just limited to a Microsoft IP
 address and they include things like cloud systems that
 are essentially publicly available, that IP filter may
 not do that much good in this specific case. Now, if you're
 using the Microsoft Agent Framework, the latest safe
 version is 1.26.0, 1.26.1 and 1.26.2 are malicious versions.
 Both are no longer available, but they were downloaded about
 a few thousand times. Actually, it says here 41,000
 downloads a week. So, it was up for part of the day
 yesterday. So, I would think a couple thousand downloads
 maybe of the malicious version. Well, and then some
 good news. The SANS Holiday Hack Challenge is live again.
 Link in the show notes. And there's also links all over
 the SANS website. So, it shouldn't be that difficult to
 find. But maybe a nice distraction sort of in
 between, you know, patching stuff and dealing with
 incidents. That's it for today. Thanks for listening.
 Thanks for liking. Thanks for recommending. And of course,
 special thanks for subscribing and talk to you again
 tomorrow. Bye.