Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

The Low Down

The Low Down artwork

Java4Shell, NGINX2Root & the Week Everything Went Down

Nov 20, 202658 minAudio
ShareXLinkedInFacebookWhatsAppEmail

Recorded Wednesday, November 18, 2026. Java and NGINX had a bad week. So did anyone trying to catch a flight. The Low Down Podcast is presented by Maze. https://mazehq.com/lowdown X: https://twitter.com/Maze_Security Ed and Matt get into Java4Shell, NGINX2Root, and why a week of outages had people asking about a cyberattack. Then we talk about what happened when Optimizely’s A/B testing scripts started serving malware. Today on The Low Down Java4Shell and “vibe uncoding” Java is everywhere. Ed gets into which apps teams need to worry about, and Matt brings up Calif’s pitch for using AI to find old code we can rip out. Cloudflare’s guidance for teams patching Java https://blog.cloudflare.com/java4shell-waf-protection NGINX2Root and what the numbers actually count Five research teams and two HTTP/2 bugs. Matt pulls apart the big exposure numbers and gets to the question a version scan can’t answer for you. Is HTTP/2 actually turned on? Wiz’s research and disclosure timeline https://www.wiz.io/blog/nginx-http2-rce-vulnerabilities NGINX security releases https://nginx.org/2026.html Why did so many services go down? An FAA ground stop, bank disruptions, and a reported NSA network outage all landed in the headlines. We go through what was known about the outages and why old systems are so hard to keep running. BBC’s outage roundup https://www.bbc.com/news/articles/c6z0kly2lll7o WIRED on how to tell an outage from a cyberattack https://www.wired.com/story/how-to-tell-if-its-a-cyberattack-or-just-an-outage/ Optimizely and the scripts running on your site A tool for testing button colors and headlines ended up serving malicious JavaScript on customer sites. We compare what Scott Helme saw with the timeline Optimizely published and ask who’s keeping tabs on the scripts we load. Report URI investigation https://blog.report-uri.com/worlds-largest-brands-infected-with-malware-via-optimizely/ Optimizely’s incident report (updated since recording) https://pages.optimizely.com/rs/947-RKE-048/images/Security%20Incident%20Report%20-%20Optimizely%20Web%20Experimentation%20-%20Nov%2017%202026.pdf Thanksgiving wrap We’re taking the week off for Thanksgiving. Thanks for hanging out with us these first six months. The Low Down is presented by Maze. LinkedIn: https://www.linkedin.com/company/mazehq/ X: https://twitter.com/Maze_Security Follow The Low Down https://www.instagram.com/lowdown.pod Ed / Low Level https://www.instagram.com/lowleveltv/ https://www.linkedin.com/in/lowleveltv/ https://www.youtube.com/@LowLevelTV Matt Johansen / Vulnerable U https://www.instagram.com/mattjayy/ https://www.linkedin.com/in/matthewjohansen/ https://www.youtube.com/@VulnerableU

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

Meacham Tower. There's a ground stop nationwide. I advise you land your aircraft. People were standing still and that sometimes is not right. Microsoft products in 74 countries and territories have been affected. I just want to come out and say we did this on purpose. Biggest banks in the U.S. They're getting hit with simultaneous outages, right? Chase, Bank of America, Wells Fargo, U.S. Bank and Truist. Even Zelle started glitching out. Apple Care, iCloud, Apple Music, Apple Pay down today. Breaking news. Everything is broken. Is somebody attacking us?

No, guys, every single one that I've looked at seems like a really boring, boring answer. We got to start somewhere, and I think we should start with... Good morning, hackers. Welcome back to the best podcast on the internet for hackers, The Lowdown, presented by Mays. Matt, I have a question. What month is it, just out of curiosity? Cybersecurity Awareness Month, part two, electric bungalow. So it's not October is what you're saying.

Because it feels like October, because we're about to talk about not only a vulnerability in Java, Java, the language sort of a la PHP, and then also vulnerability in Nginx. Are we still in Cybersecurity Awareness Month, or is it now November? What happened? The cybersecurity awareness will continue until morale improves, okay? It will continue. Like... Oh, also a couple of outages, minor outages. Okay. FAA went down and did the first nationwide ground stop since 9-11.

There was like one other in between there. And then apparently even the NSA is down. So there's a couple things happening. all the banks, and then like, yeah, Apple, Apple went down yesterday. Okay, again, I've said this on the podcast a few times. I measure how serious these things are going by like normal people in my life asking me questions. And I keep saying, I've had more normal people ask me about this than ever before. Nope, now. It's like every week, there's like a new record, the outages. I've had more people asking me about all of these outages because, and like, this is part of it, and I get it, but like, everyone is like, this must be a cyber attack.

Right. Right. Or this must be the AI agent swarm. You know, let's do it. Where's my buttons? So this is mythos, right? This is this is why the outages are happening. Right. Right. And it's like, no, guys, every single one that I've like looked at seems like a really boring, boring answer. Right. And it's like, I don't know what to tell people. But then like, why are they all happening this week? We're going to talk about it later in the episode. Yeah. We got to start somewhere. And I think we should start with... What do you want? Java or Nginx?

Let's do Java. Okay. I want to start with Java. That was totally my idea too. Yes. So, okay. I think you just went deep on this one. I went deep on Nginx this morning. We both made our own videos before we recorded this. Yeah. I think that... Okay. Like, I hear... Vulnerability in Java, and that terrifies me because it's one of the most widely used frameworks on the planet, languages on the planet, right? Like all sorts of stuff built around Java. Java runs on 6 billion devices, their tagline for the last 15 years.

Billion with a B, right? That's a few. And so, but, but, but, but I have a feeling I haven't dove as deep as you did on this one in particular because I was at cyber marketing con all week running like Volnew events and all this kind of stuff. I have a theory that this is going to be a little bit like PHP to shell. Widely used, hard to exploit. Am I right? That is completely true. So to set the stage here, right, we're all kind of in language fear land after PHP to Shell dropped.

So PHP to Shell, if you're not aware, vulnerability in the PHP runtime, right? And it was so pervasive that basically if you were on a version of PHP that was released after June of 2005, a malicious session initialization with a file upload could get RCE on a server. That sounds very bad and, like, is very bad. But as the months went on or the weeks went on after it came out, not a lot of in-the-wild exploitation was seen. The reason being, it's a networked application. It takes a lot of heap grooming, memory corruption, and really elite hacker shit to make it work.

The Java bug is actually extremely similar to the PHP bug in terms of what is required to enable exploitation. But also, not every version of Java is actually vulnerable to this. So let's kind of dive into it, right? Java 4 Shell, and first of all, the name is Java 4 Shell and not Java 2 Shell because 4 is a hilarious meme number in cybersecurity right now because of Imudra, which like props to Imudra for making that hilarious for the years to come. I saw a tweet. We got to put it on the screen. It's like there's three cyber attacks right now.

One, two, four, and three off by one. Yes. Perfect. There's two hacks and the list goes to three. I like it. It's so good, dude. Yeah, yeah, yeah. It's vulnerability in the OpenJDK, which is the Java development kit, which obviously is what you use to compile and build Java applications. The base 64 decoder. So in the year of our Lord, 2026, we have vulnerabilities in basic tour decoders. So the vulnerability is really interesting, right? If you're not aware...

Java is what is called a just-in-time compiled language. Basically, when you compile Java, it is compiled down to this thing called JVM bytecode. The JVM is basically another process that interprets the bytecode, and it's called JIT compiling, just-in-time, where it's trying to compile that code down to even more efficient native code as often as possible to make the code very, very efficient, right? You have this effectively interpreted language, and if it's being run on 6 million devices, 6 million servers, ideally you want to have special little secret food that you do to make it very, very fast.

And so the vulnerability for Java 4Shell comes down to the way that the JIT compiler optimizes the base64 operation. So when you're doing base64 decoding, you have the actual data, which is like, for example, a bunch of A's, which represent zeros in base64. And you have these equal signs, which is the padding character, which is basically going to make it aligned to the proper base64 block size. In the base64 intrinsic that runs on the CPU, so basically when the Java runtime is decoding base64, if it detects that you're doing base64 decoding operations multiple times, it'll replace your base64 operation from Java bytecode to native code, code that really runs on your CPU natively.

Now, there is an off-by-one error in the code that runs the base64 intrinsic, where basically if the input is in the form of one padding character, two padding characters, one input, one padding, one input, two padding, or three padding, it will calculate that the actual output length is zero. Then when it's doing the operation to copy it out, It will subtract one from that output length and wrap that around to max int, which will actually tell the intrinsic that it's able to copy 64 bytes to a buffer that is smaller than 64 bytes.

Classic out-of-bounds copy. What makes this even more insane is that the output character that is getting copied, you don't even control that. It's not a character that you control. You have control of one of the characters, a little control of the second character, and the third character is a constant value. It is a value that you do not control. And using that primitive, the Kalief team was able to turn this into RCE on... Hmm, what is this here? Let's see. Keycloak, Atlassian Jira, Atlassian Confluence, a J-Frog Artifactory, Jenkins, Ping Fetteri, among others.

Okay? Now... To be very clear, how are you vulnerable to this? How do I know if I'm affected by this? Well, A, you have to have a system that processes external pre-auth base 64 data. And two, the real kicker, your CPU needs to have this special CPU extension enabled. It's called AVX-5. 512-VBMI2. Basically, in the world of processors, right, they all have different hardware accelerations that allow you to do certain operations very, very fast, right?

The intrinsic in the JIT compiler for Base64, if you have it enabled, uses this instruction set to do the Base64 decoding, And the specific mask that it uses to tell the CPU how much to copy is where the vulnerability is. If you give that mask the negative one from that underflow, that will cause the out-of-bounds write. So not only do you need to process Base64 data from the user, you need to have the CPU flag enabled. And on top of that, you need to groom the heap in a very particular way per application over the network.

So is it a vulnerability that enables RCE? Yes. Is it a vulnerability that is just like command injection, get it for free, everyone can hack you? No, not at all. So, you know, bug is there, but it's not a huge deal, to be honest. So in the examples that they had, right, Jenkins, Artifactory, by the way, the same Artifactory that was the root cause of the hugging face incident, put a dollar in the hugging face jar. Can't make content without saying the word hugging face these days somewhere in the content.

Yeah, so in those examples, all of those conditions that you just said had to be true, so were they specifically crafted servers running those things in the POC that had all those things? Or is there some sort of default Jira install that has a lot of those conditions that you just listed that would be exploitable? So you would need to be running these applications that has the CPU flag enabled, one.

And then two, my assumption is that all of these applications process Base64 pre-auth. So basically the scenario is like you're a network admin, you run the JIRA Docker image on a server, that server just so happens to have that flag enabled, you're not really aware. JIRA will process Base64 for OAuth, for example. That is a vulnerable condition, right? Something like that. I didn't watch the POC. So that's how I imagine you have to set this up for it to be vulnerable.

So it's not so much these applications, even though that's where the vulnerability is. It also has a lot to do with the kind of underlying infrastructure. Yeah, so it has nothing to do with the specific application. It is the fact that the application just happens to process Base64, and the vulnerability is in the way that Java processes Base64 with this flag enabled on the CPU. Okay, cool. All right. Super interesting. If you're a network admin, maybe you're a maintainer, maybe you're a CICD guy, Khalif will put the link in the description below.

They have literally the command that you can run that will tell you if your CPU supports this instruction set, right? It's as simple as cat proc CPU info and grep for a variety of strings. Basically, AVX 512, if you look for that and you have that on with the BMI2 extension, you're vulnerable. Technically, like... Your CPU can be vulnerable. Maybe you're not processing Base64. Another way that you can actually disable this is you can use a flag within Java. Again, I'm not a huge Java guy, so I may be misspeaking a little bit. But basically, there's a flag that you can tell the Java runtime.

to not use the base64 intrinsics. It's a dash xx colon dash use base64 intrinsics that will tell the JIT compiler to not hardware accelerate base64 encoding or decoding which will make you not vulnerable to this. Either you don't have it or if you do turn it off. It's a TLDR. Got it. Got it. Cool. Amen. So I think your last bit also goes to a bit of the advice that Khalif has on this write-up, which I really like. At the end of this blog post, they're calling it vibe uncoding.

And so instead of vibe coding, vibe uncoding, using the same AI tools that they use to find a lot of these bugs to identify dead or unused or bloat code, which as we know through a lot of our vulnerability research is where a lot of these vulnerabilities wind up getting identified. So it's like, just rip this crap out. Yeah. Right. We had a Slack channel. I highly recommend this, do this and tell me if you do this right in the comments that if you do this at your day job, we had a Slack channel when I worked at Reddit called dead code society, like dead poet society. And we would just share screenshots of the pull request of like, you know, a few small green number and giant red number, right?

Like, yeah, It was like, oh, you know, plus 31 minus a million, right? It was like, oh, we were just like killing lines of code and we would celebrate that. And we all use the knife emoji in there as like, yes, nice, kill the code, kill the unused code, right? So Dead Code Society, highly recommend you do this. You're incentivizing, you're gamifying, you're celebrating the fact that you're axing a bunch of this stuff. I think that's even more important now that we're like, Vibe coding everything, right? Because now we're like definitely losing sight of like what code is in use and not.

And we didn't really have a good sight of it anyway to begin with when we were writing it all or at least including it all. But now that we're not writing damn near any of it, you know, using the tools to try to identify bloat is a very good recommendation as part of this research. Yeah. Before we leave the Caliph site, I do want to highlight. So the person that found this, my video I called him QL. I don't want to mispronounce his name. Quang Long. I'm sorry, bro, if this is me fucking your name up. He actually found the bug while trying to exploit a different bug that he found in Android and stumbled upon this bug with Codex on accident.

codecs through like 20 minutes 30 minutes of like slop bugs that weren't real like the java runtime and then it's like oh also we may have found this bug in the b64 decoder and he's like no you didn't and he looked at it and he's like oh my god it's fucking real and it only cost him it looks like 200 of token spend to not only find the bug but then to develop the first working rce Yeah, it's wild, man. So literally, my talk title, Commoditization of Nation State Hacker Skills, right? $200, you could find an RCE that this literally would have been like an APT-type vulnerability, right?

Broad RCE in a broad framework. Nuts, nuts. This team is on a tear right now. I don't know much about Khalif, right? Except for all the stuff coming out. They broke lockdown mode recently, too. Yeah, on the iPhone, yeah. Yeah, like exploit. God, it survived to restart broke lockdown mode jailbreak on modern iOS is insane. Of course, we have no details on all of that because if they did, that would be game over, right? Like we got to get new iOS versions out before we start to hear any details on that if we ever do.

Right. But insane. You know, just kudos, Khalif, that you guys are just on a generational tear right now. Um, also kudos, uh, my homies over at, uh, Mallory, not a sponsor of the show, by the way, but, uh, they were one of the sites that I was passing around when people were trying to keep, uh, ahead of this Java thing. Cause they have like a really good timeline of a bunch of different, uh, write-ups and sources as people were starting to like do this. So just like shout on this awesome timeline. I don't know. Did you see this? I put this link in the show notes. We're talking about the Nginx vulnerability?

No, no, no. Wait, what? Oh, my bad. My brain is in the wrong space. Yeah, Java Mallory thing. It's got this big timeline of all the different stuff and I just love seeing it unfold like that and how fast all of this goes and all the different sources, right? So we're not just reading it from Wiz's viewpoint on this or something like that. Yeah, sure, sure, sure. Uh, so this is super cool. So go check it out. We'll put that in the show notes. Oh, and one more shout again, not sponsored, but unless, unless, uh, cloud flare, you guys are effing killing it right now with this stuff. Also, I don't know, like the CTO of cloud flare, like, uh,

talks to us on Twitter. I don't know if you've noticed him pop up on your Twitter. He definitely pops up on mine. So if for whatever reason you watch this and you would like to come on the podcast and talk to us about how you're responding to all of this stuff at the speed that you are, I would love to talk to you about this. I think his name is Dane, right? So shout out to them because just three hours ago before we were recording this, they put out WAF rules for all of this. So even if all that stuff that you were talking about is true, and even if the patching is hard, and even if you can't turn off that B64 thing, and you do host your stuff on Cloudflare, you can put this WAF stuff.

And generally, when these big vulnerabilities come up, Cloudflare makes these things free. So that's also super cool. It's like, hey, just block this stuff at the WAF. All right. So that's one giant piece of the internet that's vulnerable about stuff. We have another. Are you done with Java? I'm good to go. Send it. Yep. Huh. Okay. So this one, we got the number two in the middle instead of the number four. So we're back to, but it's not too shell. It's too root. So it's engine X to root. I read this headline. You know, it's super funny. Um, I literally this week during all of the like dumpster fires that were going on this week, I saw multiple people on Twitter, make the joke.

Hey, you all should be war gaming and engine X zero day right now. What would you do? Because it's coming. Right. Right. And literally, they didn't know that it was coming, but the advice was just like, hey guys, you all use this. It's going to be bad when it comes. You better know, have a playbook. And boy, were they right. Can I say something real quick? Yeah. I think when we were doing the breakdown of PHP to Shell, I said that and I even said, hey, it's probably going to be a use after free because of how good LLMs are at finding them.

And guess what it is? Okay, I need crystal ball animations. I need throwbacks to the old episode. I need a black and white throwback to the episode when Ed said that and see if his prediction was correct. Or Ed, if I didn't say it, just say, LOL, no, he didn't, and then that's it. So similarly, this PHP to shell is a interesting bug pattern. It's a use after free caused by something non-fatally failing during a session establishment, but then it continues to do work with that metadata. It is likely that new bugs will be found not only in PHP that fit this model, but also NGINX, Apache, and name a network facing daemon that parses requests.

There are going to be new bugs, which is not not super great. So, uh, so yeah, it is solstice friend of the show on Twitter actually tweeted this, like this exact prediction on October 22nd. Guess when this vulnerability was found? October 21st. Wow. So you better know that somebody saw that tweet that said, you guys got a war game engine X, uh, RCEs, uh, a day after a bunch of people behind the scenes knew that there was one.

You better believe they were all sweating. That embargo broke. You know what I mean? That made me chuckle a little. I was like, Oh, someone saw that, that knew about this bug and was like, what does he know? Who leaked? Who leaked? Is the CEO of chain guard in the call? What are we doing? Oh, too soon. I'm sorry. I'm sorry. I'm sorry. I'm sorry. Chain guard. I love you. Uh, but yeah, yeah, we had to, we got to do the embargo memes here. Okay. So, uh, so NGINX to root, we've got five vulnerabilities. Okay. Um, the, the really funny part is if you go to the whiz write up about this, it was found by like this, uh, one of the main names on it is this group, uh,

Zint, X-I-N-T, I don't know how they would like to pronounce their name. But if you go to the Wiz write-up, the executive summary found by, and it's like a list of one, two, three, four, five different companies independently uncovered critical remote code execution vulnerabilities in Nginx. Two of these five vulnerabilities were found by five different companies, which means it was found by the latest model that we all pointed at Nginx, right? It's like who had the right harness with the right budget to throw the latest model at all of these modern, you know, critical pieces of infrastructure of the Internet and Nginx and two of these RCEs fell out.

That's literally all that means, right? I don't know which one it is. Probably Sol56. Does that timeline line up? Maybe. And then Zint found three other vulnerabilities. So that tells you harnesses still matter. The people running these frontier models still matter because some of these teams found additional vulnerabilities than others instead of just like YOLO, Sol56, please go find me, Nginx, RTE. Go now. Yeah, exactly. So it still matters. Talented researchers, good harnesses, this stuff still all matters.

Also, the ability to sort through slop, right? The ability to know what is a real bug, what is exploitable, and that matters too. You can have a good harness, and it's still going to throw nonsense at you, right? Totally. You need to figure out what's legit. Yeah, you got to gaslight your model. It's going to gaslight you right back that it found something, right? And so it's like the gaslighting games, the Hunger Games. I need gaslighting games of the AI models to figure out what's actually vulnerable. All right. So this one's obviously crazy, right? If you're listening to this podcast, you probably know what Nginx is. But if you don't, it's like one of the most widely used web servers on the planet.

It's like Nginx and Apache, right? It's like what's out there. Estimated stats, 225 million websites are powered by this. Nginx also very regularly is the thing that sits in front of your web application, right? So like I've used it mostly in that method, you know, like load balancing, reverse proxy type stuff where it sits in front of your application. So even if like, oh, no, I run on, you know, XYZ stack or like the Java stuff or like whatever, there's likely an Nginx box sitting in front of the web application no matter what stack you run.

Yeah, literally every web app that I've made is like Django or something, but I always put Nginx in front because CertBot for TLS can interop with Nginx. You do CertBot deploy on Nginx and you do unencrypted back to your backend, right? So it's like it's everywhere. Exactly. The other word that you just said, super important, because part of the why this matters, right? Remote code execution on that Nginx box. is being able to intercept information that is being proxied to that. And you might say, well, Matt, in low level, I have TLS.

I don't need to be worrying about interception. But like what Ed just said, a lot of times, a lot of times, what you're doing is terminating that TLS session at the Nginx box and then passing the unencrypted traffic back to your actual application to do whatever between your app and your database and all this kind of stuff, right? Not to say that's all the time, but that's very common, right? Super, super common. And then even if you didn't, it doesn't matter because you have remote code execution in the critical path to the application. So you can actually change the code of the website that's being served to the browser.

Guess where the key material is. It's in that process. It's game over. You could do, like, the attack scenarios are, like, endless besides the man in the middle. I'm just saying TLS is not exactly like, oh, you're good now here because of this particular bug. So it's only rated a 9.2 CVSS instead of a 10. And that's because it, I think probably likely that the docking is because it requires HTTP 2 to be enabled on the Nginx server. So that's the one part of this story.

This story broke two hours ago. Yeah. The one part of the stats that I do not have is that, the HTTP2 stats. It's harder to fingerprint on the Shodan type stuff. 225 million websites was the estimate of just Nginx installs. On here, it also says 98% of Shodan identified Nginx servers, which I think was about in the 50 millions, was Shodan visible boxes.

98% of those 50 something million were running vulnerable versions of this. So basically anything that Shodan can see is vulnerable to this for the most part in broad swaths. But based on version number, that does not tell us if HTTP 2 is enabled on all of those. I think it's not, you said it's not by default. Yeah, so H2 and H3 are not default behavior on Nginx. Sure. But, uh, but it's super calm.

I know it's super common. I just don't have the stats to tell you right now on that particular thing. I, if you're listening to this, if we get this out, you know, if we turn this around, uh, fast enough this week and you're, and you hadn't responded to this yet, I would assume, I would just assume that, that you're running one of these versions and that it's, uh, vulnerable. Just act like that. The good news about that particular caveat, right, whether or not it's widely exploitable or not, is that if you can't patch for whatever reason, right, that's all you need to disable in order to do it.

So the fix for this one, like, fits on one screen, right? On this Nginx to root page has, like, the what if I can't patch section. And, like, on one screen, it's got two commands that's just like, okay, HTTP2 off. And it's like, if you could turn that off, if for some reason that's easier for you than upgrading to the latest version, that would stop this from being exploitable too. It's also super easy to test if you're exploitable. So this is a little bit different than the Java and the PHP in its widespreadness.

It seems easier to exploit from what I can tell. There's no active exploitation that we know about right now. But the researchers have not released a POC, but they've released videos of the POC. that this is easier to exploit than the PHP and Java situations. And it's easier for you to test yourself if you are impacted. It's also just one command that you can run to see if you're impacted by this. You run a second command to disable HTTP2 or to update, you run that same command, and you can test whether or not you're vulnerable to this.

So this should be a relatively easy fix compared to all the other crap that we're doing right now. You want to talk about the bug? Like the use after free bit? Yes, we haven't seen a pop, we haven't seen the exact reasons why, but if you read the CBE itself, the CNA being F5, Let's see. When Nginx Plus or Nginx Open Source is configured to use H2 for client connections, a crafted request can result in a use after free in the worker process if this process is running low on worker connections.

I want to highlight that piece. It's an interesting... I think illusion to a pattern we're going to start to see as LLMs get better at pattern matching and doing VR, right? This is a similar code pattern to the PHP to shell bug, right? The PHP to shell bug was if you can cause the allocator to fail during a session upload, the session pointer will reference freed memory, right? It's the same thing. If you can cause a failure state to occur... Because all the workers are being occupied and you're trying to maybe spawn another worker to handle another H2 chunk, that causes a failure that references freed memory, right?

So that's pretty interesting. We don't know why, but it sounds like some kind of teardown that occurs when you need a new worker but you don't have one. You know, you tear down part of an object and another part of the memory references that dead object, right? One thing, I'm a huge asshole when it comes to verbiage or phraseology in CNAs. Literally the next line, or in CVEs rather. Additionally, attackers can execute code on systems with ASLR disabled or when the attacker can bypass ASLR. Oh, so A or not A, which means one?

Like, what are you talking about, dude? I hate that. What I think they should probably say is, like, the researcher has shown that they can bypass ASLR with this, right? Similar to PHP to Shell. You have to do kind of like a leak via the UAF to get the ASLR info out and then use that to calculate the memory map. Interesting verbiage here for sure. Yeah. Either way, I think it's going to be as complicated, potentially if not more complicated, than PHP to shell to exploit. Therefore, we will see probably a similar amount of exploitation of this, which is like low to zero.

And unlike PHP, to your point, you can turn this off with a simple flag in your sites-enabled config. So go turn off h2 and h3, and you're good to go. Yeah. So good write up, good work here by everyone who, like I said, had the budget and the latest model to point at things like Nginx. I'm glad that that's our new reality, right? Is these new models are going to come out, they're going to be better at finding specific types of vulnerabilities in specific types of code, right? There's all these nuances to each model that comes out, right?

And right now at this point, we have enough security research teams that have gotten on board with this and the harnesses and the vulnerability research, you know, budget and divisions and teams focused to grabbing that new model and pointing it at stuff that's super important and critical load bearing things like Nginx or like Java, apparently, even though that one, like you said, was an accident. You know, we're pointing the right tools at the right things. This is why, since Mythos and its Project Glasswing velvet rope invite-only situation, I've been advocating to get these things into the hands of these types of people.

Some of these researchers were not in Project Glasswing. They're not major Fortune 50 companies that are doing this. These are talented vulnerability researchers with the ability to create these harnesses and, like you said, understand the results. And they're finding... very critical things and helping get them fixed. And by the way, look at us. Congrats, cybersecurity. The embargo didn't break on this one. We have a patch. Like, this would have been a very different conversation if this leaked October 22nd and there was no patch. And we just had to go out there and be like, I don't know, everybody, go figure out your Nginx stuff.

Well, in this case, you could just disable HTTP2. But... Maybe some people can't do that. Maybe you absolutely need a patch, right? So, yeah, this one was cool. And like I said, even not major companies that were invited to that velvet rope glasswing situation could play by these coordinated disclosure rules and, like, do for the better of the Internet type stuff that we just did here, right? Matt, do you know who else is good at finding vulnerabilities in code? The sponsor of today's video! Hey hackers, before we keep going, I wanna take a quick break and say a quick thank you to our sponsors.

Today's Lowdown episode is sponsored by our friends over at Maze. Maze has spent the last couple of years using AI to find and remediate vulnerabilities in the cloud. And now they've released Maze Code, AI agents that find and deeply investigate vulnerabilities in your code. Now we all know the deal. Software composition analysis tools and SaaS toolings are all historically really bad at just making more noise. They find issues that may not actually be real and just create a lot of problems for your engineers to deal with that don't solve real security vulnerabilities.

I get it, we've been there. maze code is a really cool tool because maze code actually knows how your code works maze code investigates every finding with the context of how your code is actually deployed sure you're using a version of lib xml2 that's vulnerable the actual code path to hit that vulnerability is not exposed in your docker image maze code can find that and won't show you that that vulnerability is a big deal Every vulnerability that Maze code reports comes with evidence that it's real and a way to fix it. Check out Maze at go.thelowdownpod.com slash Maze.

That's go.thelowdownpod.com slash Maze. Thanks again, Maze. Let's keep going. And we're back. Hey, welcome back, everybody. How are we doing? We're professional podcasters, bro. Throwing it out to the sponsor, getting it back. We're so good. So good. What percentage of podcasts make it past 20 episodes straight? Like, we did good. We did good, dude. I think we're doing all right. All right. So it's 2027. It's going to be our year. It's going to be great. So I've been resisting saying this as part of the whole Nginx story because it would have transitioned us directly into the new story.

And we needed to finish the Nginx story before we could transition into all these outages. But... Like one of the things I keep bringing up and we both just kept bringing up in these two stories was if it's difficult for you to patch for some reason. Right. That was like a sentence that I just kept saying. Right. With NGINX. Okay. We'll disable HTTP2. why would it be difficult to patch? Like if you're listening to this and you don't work for some like megacorp or anything like that, and you're just like, well, I have Nginx, so like upgrade, right? Done. Cool.

Cool, right? That's not how life works. Yep. When you run in like a network the size of France that if it goes down, people's paychecks don't clear, right? You know, so like you can't just be like, cool, upgrade, you know, lol, no vuln anymore, right? So this is what I think is going on with all these outages, right? I like... Should we list the outages first? I was going to say, yeah, can you go into, like, you say these outages. There are several in the plural these, so give me a few. Sorry, I got ahead of myself, everybody. We have an article from the BBC headlined,

Hacks, AI, or just overdue updates. Why have there been so many outages? We have escaped the echo chamber drastically right now, right? We've got major news, like websites, talking heads, constantly saying, you know, what's the deal with these outages? I saw press releases in the White House asking if this was cyber attack related or AI related. And they had to go, well, actually, around here we call it SI. It's like, that's not the point. Can we get – the planes aren't taking off.

Why are we – The Kalief.io Java for Shell story made New York Times talk about heap grooming. Like, we are out of the bubble, dude. It is happening in real time right now. Wild. Machine speed. I got to see that. Okay, cool. So outages. It started with Microsoft, at least this little stint that we're talking about. Yeah, we're not in Black October anymore. That was a different stint and pile of issues. The Microsoft outage was kind of the kickoff of this, right? We made the joke on our last podcast, Microsoft 364, am I right?

And they were down for the better part of a day. We still don't have a lot of information about what happened there. right? Like that's kind of wild that like some sort of transparency report hasn't come out about that one. But there was some like theories that maybe that caused a lot of other ripple effects because Microsoft is so load-bearing for so many organizations. Right after the Microsoft outage, like the next day, I don't know, time's fake. This is a few days ago. The FAA grounded all flights, like we mentioned at the top of the episode.

People are like, was this AI? Are we under attack right now? I saw that it started in Chicago and started to spread out. Is this a worm hitting air traffic control systems? Whatever, right? Then... a bunch of the biggest banks in the U.S. went out, like yesterday or the day before, right? Then the NSA's confidential network, NSA net, went down. And the stuff started going about that. Then, like you said, the New York Times talking about Java bugs, right?

So all of the media is like, what is going on, all of these outages? Apple had an outage. Like the App Store and Apple Wallet went down for a little bit this week. The meme was like the Bart Simpson say the line thing. And he's like, there's no indication of a cyber attack. Like, yay! Yeah. I did the breaking your mom is down. No suspected cyber attack. It's just like, it's just constant, right? And so, yeah. Right now, it seems like all of these are very, very boring IT issues.

It just seems like they're all happening at the same time. So that begs the question, is this a coincidence? Or is this some sort of giant government conspiracy? And there is actually some sort of like AI thing going on or cyber attack going on. And we're all just getting fed, you know, some story about boring IT things. And... I come back to, and this is like my take on this, and I'm excited to hear yours, right, is I think the most plausible explanation of all of this happening is no, it's not a coincidence that this all happened at the same time.

It's a ripple effect of us identifying more issues right now, right? So it's not like, oh, AI caused all this stuff, but we are finding more bugs, not just security bugs. We're finding more bugs in general. And so... Patching and upgrading are risky behaviors. This is why like enterprises do code freezes around Black Friday or around like big important days or when a bunch of their staff is going to go on vacation because you do a code freeze because upgrading is risky and like things break without you realizing it.

It's hard to test every scenario that's like live in production, right? So once you do go live, there's risk of this. So I think that it's pretty plausible that... That we're at this point, the holidays are coming up. People are like scheduling all these upgrades from all of these things that we've been identifying at a faster rate than ever right now. And like we've squished and compressed a bunch of our issues to right now. What do you think is going on? Yeah, I think I agree with you, right? I mean, there are a lot of... No, come on.

I need the tinfoil wizard hat. I need you to tell me that this is North Korea. I need you to tell me this is Iran. Well, so, okay, okay, okay. When this first started to happen... Right. When like Microsoft went down and then the FAA went down and then airlines went down, I was like, OK, maybe this is an Iran thing. And then we had like the news of like the dirty bomb threat going on. So I'm like, OK, maybe this is Iran, you know, trying to flex if they can act like a nation state, too. But then, like, the nothing ever happens meme kind of started to go off in my head.

And I'm like, yeah, but, like, nothing happened. Like, no planes flew out of the sky. There was, like, a pool chlorination system that, like, over-chlorinated a pool in France because it depended on Microsoft Outlook. So, like, that's bad. But, like, other than that, nothing really, like, occurred. So I don't think it's a nation-state thing. I think I agree with you, man. I think it's literally just the... The technical debt or the security debt habits that we've kind of built up as an industry are starting to catch up with us. And now we have to come up with new ways to undo those habits.

And, yeah, I think, like you said, patching is risky. Updating is risky. And so you have organizations like the FAA, you know, Citibank, maybe the NSA. They update that one router that's been sitting there for 20 years no one's ever touched. And then suddenly it takes out the national grid for the radar. I don't know what's going on, but I think I agree with you that it's just the patching of these bugs is causing ripple effects that we can expect. Yeah, and I do like the theory that a lot of this is happening right now because the holidays are literally next week. And they know everyone's going to go on vacation, on leave, whatever.

Everywhere I've ever worked has code freeze starting next week through the end of the year. It's like, nope, we're not doing anything, by the way, like life support or major security issues, exceptions or whatever. The joke I made was that that load marrying it guys playing wow forever. And he's, he's just been out of office for too long. He's already a level 55 shaman. It's only been out a week. Uh, and, uh, that's why, but that was the joke I made. But, um, yeah, I, I really don't see, I don't see any evidence in any of this being like cyber attack related. And people were in my DMS all week, all week, man.

Like, where's the video on this? Where's the video on this? I was like, where's the video on what? There's no cyber. There's nothing to talk about. It's like, yep. It went down. What else do you want to Tons of other creators that were making videos about it going down and like, oh, is it a hack? We don't know. And it's like, well, that's not useful. Like, you know, and I'm just trying to not add to the noise, everybody. Okay. Like, you know, is this AI? Is this like whatever? I'm like, there's just noise. It's just noise. I don't know. And I got misinformation and then contributed to it. Like, Polymarket said, FBI takes possession of the router in question from the FAA.

And I retweeted that. And then the FAA tweeted, and they're like, hey... Or no, the FBI tweeted, and they're like, hey, we didn't take any routers from anybody. That was made up. Oh, wow, I missed that. Yeah, that's crazy. I mean, it's nuts. I'm a little, like... I will say, I'm not a big conspiracy guy, but, like, there's a little bit... I mean, I start to feel like the Charlie, you know, It's Always Sunny gif, right? Like, there's too many things going on, right? I mean, I'm in text messages with, like, much more important people than you and me, and, like, that's not a flag.

I'm just, like, in some groups, and, like, these people are, like... also kind of scratching their head and like starting to ask questions. And it's like, okay, hold on, right? And like a lot of people don't believe in coincidences. And that's why I wanted to say, like, I don't think it's a coincidence. I think it's because of like, everything affects everything, right? And like, we've got these super bug finding tools. You and I talk a lot about security bugs, but that's not the only kind of bugs. Yeah. But AI is finding in a lot of this stuff, right?

It could just be like, hey, we're trying to optimize this, or we're upgrading this, or we've got a lot of pressure from leadership to be running frontier models to do something, save money, fix bugs, modernize, whatever, right? And the pressure to move fast and break things, guess what? We move fast and broke things. I think the reason why the trigger pull is so light on at least internally our industry freaking out, At least for me, I think there's this general awareness of how fragile the system is, how fragile everything is.

And if you apply LLM bug finding to that, you become aware of the system where like, hey, everything is kind of popsicle sticks and bubble gum. And also nation states probably have bugs in all of it. So then you're kind of like almost waiting for this scenario to occur. I think back to like the Volt Typhoon, Salt Typhoon, Telecom intrusions where it's like, oh, you know, Chinese threat actor hacked into Verizon and T-Mobile and AT&T and Sprint and, and, and, and, and. And so you have these APTs that are probably like positioned in some of these networks.

You're almost kind of waiting for this day to happen where you're like, oh, yeah. Microsoft went down and the FAA and the NSA and banks and Java. Like, it's kind of like, oh, duh. You know what I mean? So you almost want to believe it because it feels like it's possible. But that being said, still no indication of cyber attack. We have no idea. It's just I think that's why we're so ready to go that route of doomsday because it's the industry is like set up for it. You know, it's all it is also I'll give it to the conspiracy theorists. It's a little weird that they're literally all unrelated. Yeah, right. I agree. The fact that none of these things helped cause the other one of these things or that they don't share a root cause is strange, right?

Like super, super strange, right? I get it. I think Sherrod over at Unit 42, previously Microsoft tweeted it really well. She said, it's interesting that people in threat intelligence with deep expertise in threat actor activity like Russia, China, Iran, North Korea specialists easily accept that the recent issues are human errors made on fragile systems. while people without threat intelligence backgrounds are shouting OMG cyber attack. I think that sums it up really nicely, right? It's just, yeah, it's just super weird that these things, like when I read, like you texted me the NSA thing, I was like at a conference and I was like, this could be good.

Oh, shit. okay, I guess I'll make a video now. And fool me six times, shame on me, right? I guess I'll talk about something. But yeah, it's just duct tape and popsicle sticks all the way down, man, is really what's going on. And we just started blowing harder on this fragile tower. The quote that I like is that we've had decades of tech debt forever, and we all know this. If you work in technology, you all know this. Nowhere that anyone's ever worked is like, Modern, got it, done.

My system is good. No one says that. Nope, nope, nope, right? But certain tech debt is worse than others. And guess what? In this age right now that we're in, interest rates just went through the roof on tech debt, right? And I think this is part of that. It's not the vulnpocalypse. Well, it's kind of vulnpocalypse. It's fucking Java and NGINX, right? But it's not the hacker-pocalypse, right? Or like- The cyberpunk-harmoner, if you will. Totally, right? It's not that. But- the interest rate's going up, meaning we're going to have some outages. Guarantee, by the way, Nginx upgrades or Java upgrades are going to take some stuff out.

Whether or not it's big swaths of things like banks, right? Because they have redundancy testing. But someone's going to stub their toe trying to rush around and patch these massive... When Nginx is going to find out that the other end of that connection depends on HTTP2, and it's like, uh-oh. 100%. 100%. It might be like some homies blog and not like, you know, Bank of America, right? So like, who knows? So this one's wild. I just like, I kept like holding my tongue transitioning from the Nginx story to this because I kept saying patching is hard and risky. And I was like, wait, we got to finish Nginx.

Yeah. All right. So what's the other, this other bit that absolutely took up a ton of air this week? Because it's like every big name company, SaaS company was listed as like hacked basically in the headlines for people on this one. Did you see this one? Optimizely? I saw floating around on Twitter. I couldn't give you a ton of details, but it sounds like it affected a ton of downstream companies. So please enlighten me. Yeah, so you and I both had to figure out what Optimizely was first, right?

So it's basically a JavaScript front-end framework that you can include on your website that's code that you don't write. So this is technically, now, let's call it, everybody put a dollar in the jar, supply chain attack. Supply chain! So if you remember these from way back in like, I don't know, what, September? Pre-Black October and vulnerabilities and everything, what was the other topic du jour was supply chain attacks, right? With Team PCP and... Shy Allude. Shy Allude, NPS. NPM worms, PyPy worms, rust crates and all this kind of stuff.

This is in that vein, right? But instead of hacking GitHub accounts and getting into different NPM packages, this Optimizely company just got hacked to themselves. And then the JavaScript that they were serving to their customers all of a sudden started serving malware, right? So this JavaScript that everyone had included on their site, if you've never done this, you don't actually include the whole JavaScript that does all of the functionality. You include a line on your website that...

in real time calls out to the other company and brings in that JavaScript at the moment that someone's browser loads it. So this, if you want like a Matt J history lesson lore dump, you can go Google my Black Hat talk in 2013 about building a botnet on an ad network. And this is exactly how I did it. I got an ad approved on an ad network that included JavaScript. And then I changed the JavaScript on my server that it was pulling to be malicious and but the ad was already approved and out there and it didn't matter and that line of code never changed in the ad anymore i could just change the code that it was pulling into the browser at that time right uh not to date myself on my security research back to 2013 that was my first black act don't oh was it were you in my talk i may have been there actually no i don't know

So it's a web app. You're cooler than web app security. I didn't know shit back then, but anyway. Web apps are not very... Low level. In the channel. So, yeah. So big names like Zoom and the BBC and HSBC and Equifax are on the headlines of this. So what this means and why this is scary is these massive websites that you use and trust. You go to Zoom all the time. You go read the news on BBC. You expect that website to not serve you malware.

And for a period of time this week, they were serving malware on these very legitimate sites. because of this malicious Java includes. So there's 178 organizations that used Optimizely, which by the way is like an A-B tester. So it's like you can dynamically generate this JavaScript on the page that does like an A-B test of, I don't know, various things. Does this image or this image sell more t-shirts? Like which one? That's the whole thing. Literally, right? So pulled this in and what did the malware do?

Steal crypto. Of course. Of course. Of course it's still crypto. So it was like card skimming. It was like crypto address swapping in page text forms, like from the clipboard and like outgoing requests. So that's like the only confirmed loss. That's how it happened. It was like... making a transfer to something and it like swapped their crypto wallet address for the actual one. So I think the only confirmed loss is about $7,000 from this. This could have been a lot worse, right? But it looks like it was also throwing like Karuna and Darksword exploit kits.

Yeah. Which is insane. I was getting there. So like cards, no, it's all good. So card skimming crypto was like the main things that like that's the only confirmed loss. But then like, but as the research was like unfolding on this, it even says like added partway through the day. So this compromise happened. The attacker was testing what they could get away with. And then they were adding new malicious features to it as it was going. And one of the later features was throwing Karuna and Dark Sword. I have a video on my channel where I interviewed the researchers that identified Karuna and Dark Sword, if you're not familiar with those.

But their iOS exploit kits that were found in the wild earlier this year. that previously would have been like 20 plus million dollar exploit kits that were now just like sitting on random watering hole websites, like just hacking anyone who came across that website, which is not how iOS exploits have ever been used in the past. They are very targeted because they are very hard and expensive to develop historically. And so then here you go. Karuna and Darksword kind of resurged and were part of this in a hidden frame.

It was depending on certain iOS versions, right? Most of them were older. Some of them were slightly newer. But yeah, this is nuts to see the iOS exploit next to the crypto scamming and card skimming again, right? These are not... We're not used to this as an industry. That level of OS exploit being on crypto. The only money stolen from this whole campaign was $7,000. And they were using a previously what would have been a $20 million exploit as part of this attack.

It was just insane. It is interesting that the world of go to a bad website and get hacked by a nation state is still alive and well. You know what I mean? You think like, oh, with all the... you know, looking at V8 we're doing and the fuzzing of this and the hardware mitigations we have on iPhones that would go away. But it's like, no, bro, you could still get popped by going to, I mean, fucking BBC.co.uk. You can get hacked, you know, insane. Yeah. So, you know, zooming out on this one to not just be like breach porn or vuln porn, like we say, right?

I think the lesson learned that we want to walk away from this is like part, you know, that dead code society thing we talked about and like unused code, add this to the list. What are you including in your JavaScript front end that like this A-B test thing that like, what, someone from marketing or like marketing ops, like growth ops, like this is like in that world of like automation around optimization of your stuff, like through this on your website? Did you know it was there? Right. Like, did the security team at BBC know that this was there and evaluated it?

Did they have any sort of way to monitor if optimizely was changing the code that was being pulled in dynamically at, you know, runtime in the browser like this, we, you got to have answers for this and you got to evaluate like your attack surface, like has always included this stuff. But it's time. Tech debt interest rates, right? It's time to evaluate what are you actually pulling in. The other follow-up from this is literally just this morning, we got an incident report out of Optimizely about this.

I haven't actually got to read the full report, but the researcher who additionally... found it, said that they confirmed infections outside of their timeline. So they state that first infection was like, oh, 700 hours or whatever, but they have it going like maybe three or four hours even before that. So even the researchers like questioning this whole incident report. So we might actually learn some more about this as it goes, right? So it seems like maybe incident response isn't exactly going super well over there.

I don't know if... I don't know what's going on. So maybe a moving target of a story, but definitely a broader thing to be concerned about when it comes to these JS includes. Interesting. Love it. All right, man. Black October behind us. Whatever the hell November is in front of us. I am not looking forward to the headlines during the holidays. The holidays are always when hacks happen. And the hacks have been happening off the holidays. I don't want to know what's going to happen on the holidays. This is when actual nation states kind of wake up and start to poke at us.

And that's not what's going on. Everything that we talked about today was security researchers, you know, finding this stuff, not nation state hackers or cyber criminals, right? So crystal ball for future podcast episode edits for me is that Thanksgiving and Christmas are going to be pretty spicy over here in the U.S. Yeah, I think the winter is going to be cold. And I think other nations know that. So I think the power of greater others may see some fun as well. So we'll see. Very Game of Thrones of us, right? Yeah, I know. The winter is coming.

Winter is coming. Hey, but real talk, guys. Thanksgiving is coming. We're going to be off next week due to the Thanksgiving holiday. I do want to say we are quite thankful for you all who listen here on YouTube and on Spotify and Apple Podcasts. We started this back May of this year, and we really could never have imagined it would have grown to what it is right now with only six months behind us. So seriously, thank you for listening. Thank you for hanging out with us. 2027 is going to be a huge year for us. We just want to express our gratitude for your support, our sponsor's support, and everyone else who's here to hang out.

It's been a good time so far. And thanks to Mays, our sponsor, for doing this. They believed in us from episode one, literally a launch partner. Couldn't have imagined that. Turns out, you know, our intros are polarizing, but I freaking love them. And it turns out editors cost money to do that kind of stuff. So thank you, Mays, for enabling the production values to this so that we're not just publishing two guys, you know, talking on Zoom. Like, you know, we're trying to put out a good podcast for you guys. I couldn't have said it better myself. Super thankful.

Thankful for you, Mr. Low Level. And thankful for the audience. And this has been super awesome. Do you have any traditions? I know we're wrapping it up. Do you have any Thanksgiving traditions? I have one in particular that's like a special stuffing recipe. That's been in my family for quite a while. It's like a sausage stuffing, like mashed potato sausage, not like a bread stuffing. But it's kind of funny because then some of my family wants the bread stuffing and some of my family wants mashed potatoes. So we kind of have like mashed potatoes, mashed potatoes and like sausage and then bread stuffing.

So there's like a lot of sides. The whole gradient, yeah. Yeah, yeah. And then also it's like, we kind of pride ourselves in being the open door. So we collect strays. So every Thanksgiving, it's like, who's out of town? Who just moved here? Doesn't have family, whatever. So I always have a gazillion people in my house and I love it that way. And it's awesome. I've had Thanksgivings where folding tables were snaking through the house because people couldn't fit at the dining table. It just feels good. It feels like the start of the holiday season and collecting family and all that kind of stuff.

Yeah, we do kind of the same thing. So we alternate between we host and then my mother-in-law hosts. So this year we're hosting. We're hosting 20 people, which means we're doing that. We're doing that where we have like our main dining room and then like folding table, folding table out into our reading room. Meal, you know, food-wise is pretty traditional stuff. Turkey, mac and cheese. I make like a mac and cheese with a meat sauce kind of thing. It's kind of like almost like a hamburger helper sort of vibe, like real pasta and stuff. And just hang out and have good food, man. It's going to be great. I got really into, you know, I'm a millennial white male with a beard, so I got really into smoking barbecue for quite a while in my life, smoking meats, smoking meats.

And so, yeah, of course, for a few years, I was smoking the turkey and people liked it. But the smoked turkey hit or miss, honestly, I kind of like the traditional oven one better, unless I will actually sometimes like cater some turkey from a local Austin barbecue restaurant. Even if we cook a whole bird, It's like kind of fun to bring in some local Austin barbecue, but I do make, and this is always a hit, uh, is a, uh, cranberry chipotle barbecue sauce to go with a smoked Turkey. So I use like whole cranberries and whole like chipotle peppers. And I make barbecue sauce and like immersion blend the whole thing.

That sounds really good. Oh my God. It goes real well with smoked Turkey. It's like, I don't know. So especially if family comes in from out of state, it's kind of like, Hey, we're going to do a little Texas thing here for you, you know? There you go. So that's, uh, that's super fun. Cool, man. Alright y'all. Sweet. Well happy hacking, happy Thanksgiving, and we'll see y'all in two weeks. Peace.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. This SSH Bug Accepted Wrong Passwords, AI Exploits, Tiny Keys & a Surprise Guest

    Nov 17, 20261 hr 11 minAudio

    Recorded on Friday, November 13, 2026 The Low Down Podcast is presented by Maze. https://mazehq.com/lowdown X: https://twitter.com/Maze_Security 257 A's should not get you into an SSH account. John Hammond joins Ed and...

    Credit the people on this episode

  2. Binance's Fake Interview Hack, Railway's AI Incident, Hacked Freezers & a Critical Linux Bug

    Nov 6, 20261 hr 6 minAudio

    The Low Down Podcast is presented by Maze. https://mazehq.com/lowdown X: https://twitter.com/Maze_Security The "signup test" was probing for Kubernetes secrets. Ed and Matt get into Railway's AI testing incident, hacked...

    Credit the people on this episode

  3. Black October Recap, OpenCode’s Stolen API Keys, AI Distillation & the LiteLLM Supply-Chain Attack

    Nov 3, 20261 hr 24 minAudio

    Welcome to The Low Down, the best show on the internet for hackers Black October has real bugs, stolen keys, and a lot of very loud headlines. Ed and Matt sort through PHP2Shell, AI security claims, and the difference...

    Credit the people on this episode

  4. Cybersecurity Experts REACT to the php chaos... (and MORE)

    Oct 28, 20261 hr 28 minAudio

    Welcome to The Low Down, the best show on the internet for hackers Listen on Spotify! https://go.lowdownpod.com/spotify The Low Down is presented by Maze. LinkedIn: https://www.linkedin.com/company/mazehq/ X:...

    Credit the people on this episode

  5. Cybersecurity Experts BREAK DOWN the shocking OpenAI Spying Incident, Crypto Backdoors, AND MORE

    Oct 19, 20261 hr 17 minAudio

    Welcome to The Low Down, the best show on the internet for hackers Listen on Spotify! https://go.lowdownpod.com/spotify The Low Down is presented by Maze. LinkedIn: https://www.linkedin.com/company/mazehq/ X:...

    Credit the people on this episode