This transcript was generated automatically, and has not been manually verified. It may contain errors and omissions. In particular, speaker labels, proper nouns, and attributions may be incorrect. Treat it as a helpful guide rather than a verbatim record — for the real thing, give the episode a listen.

---

GRAHAM CLULEY. So if you can secure that, you're going a long way to preventing yourself from becoming the next cybersecurity headline. Beautifully said.

---

TRICIA HOWARD. Beautifully said.

---

GRAHAM CLULEY. Thank you.

---

TRICIA HOWARD. It's almost like you have a show that you do for a living.

---

GRAHAM CLULEY. I wouldn't call it a living, but yeah.

---

TRICIA HOWARD. Okay.

---

UNKNOWN. Smashing Security. Episode 444: We're Sorry, Wait, Did a Company Actually Say That? With Graham Cluley and special guest Tricia Howard. Hello, hello, and welcome to Smashing Security Episode 444. My name's Graham Cluley.

---

TRICIA HOWARD. And I'm Tricia Howard, AKA Tricia Kicks Sass.

---

GRAHAM CLULEY. Tricia, hello. You're new to the show. Where have you come from?

---

TRICIA HOWARD. I am so excited to be on here. I have been a fan of this show for such a long time. It's an honor.

I work at Akamai Technologies. I represent the security intelligence group here and I work with our hundreds of researchers globally. They break stuff and I help them write about it.

---

GRAHAM CLULEY. Oh, cool. Now, where do you get a nickname like Tricia Kicks Sass from?

---

TRICIA HOWARD. Ah, so of course, security as a service.

---

GRAHAM CLULEY. Oh, yes.

---

TRICIA HOWARD. There was a vendor that came in one time at my first job, long, long time ago, and they had said something kicks sass, and I thought that was so funny that I stole it, and it has become my name ever since.

---

GRAHAM CLULEY. Most of the best things are stolen, aren't they? That's basically how the British Empire was born.

---

TRICIA HOWARD. I prefer inspired by, but I definitely stole this one.

---

GRAHAM CLULEY. Ah, okay. Now, before we kick off this week, let's thank this week's wonderful sponsors, Vanta and Horizon 3 AI.

We'll be hearing more about them later on in the podcast. This week on Smashing Security.

We're not gonna be talking about how a simple security flaw in WhatsApp exposed 3.5 billion phone numbers. You'll hear no discussion how British hacker must repay £4 million after hijacking celebrity Twitter accounts.

And we won't even mention how Cloudflare experienced a massive outage taking down large chunks of the internet. So Tricia, what are you going to be talking about this week?

---

TRICIA HOWARD. I'm going to be chatting about a new malware campaign that is targeting hotels and their guests through semi-legitimate means.

---

GRAHAM CLULEY. And I'm going to be asking why more companies don't say sorry. All that and a featured interview with Snehal Antani of Horizon3 AI coming up on this episode of Smashing Security.

Right, we've got a chance now to thank one of the supporters of this week's podcast, Horizon3 AI. You can't defend what you don't see, and that's why Horizon3 AI created NodeZero to continuously test your network the same way real attackers would and built to help you prove your defenses work.

Traditional pen tests happen once a year. They're manual, they're expensive, and they're outdated the moment they're done.

NodeZero changes that by continuously testing your environment. With over 170,000 pen tests completed, NodeZero doesn't just find vulnerabilities, it proves how they can be exploited safely.

From Active Directory tripwires to AI-driven attack paths, you'll see your network the way an adversary does and before they do. Join thousands of organizations who've moved from reactive to continuous security because the best defense is understanding offense.

Visit horizon3.ai to get your autonomous pen test demo today. That's horizon3.ai.

And thanks to Horizon3 AI for supporting the show. Now, chums, chums, and you, Tricia, you know what really gets my goat, Tricia?

---

TRICIA HOWARD. WhatsApp?

---

GRAHAM CLULEY. No, not WhatsApp. What really gets my goat are companies who won't say sorry. Right?

---

TRICIA HOWARD. Sorry?

---

GRAHAM CLULEY. Yeah, sorry, exactly. I know you say it like you've never heard the word.

Sorry, you just never do hear the word, do you, these days from these companies?

---

TRICIA HOWARD. No.

---

GRAHAM CLULEY. They are terrible at apologising. They're terrible at apologising when they take advantage of your personal data, for instance, the privacy controls or, oh dear, we've accidentally turned on this new feature and forced you to log in if you don't like it. But in fact, we're already scraping all your information for our artificial intelligence large language models.

---

TRICIA HOWARD. But they really care about security.

---

GRAHAM CLULEY. They really care about security. They really do. And they really care about privacy as well.

Or they won't say sorry because they've just suffered a cybersecurity attack. They are terrible at admitting fault for that.

They will stall. They will waffle.

They will hide behind phrases like "out of an abundance of caution" when notifying you that your firstborn child and your house are no longer your property. I think it's a mixture of fear and ego, but mostly it's those— was it Wormtongue in Lord of the Rings?

Used to whisper in the ear of the king of wherever it was. I'm sorry, all the Lord of the Rings fans who remember who it was.

Anyway, the lawyers are whispering in the ear, aren't they? They're saying, admit nothing, don't own up to anything, because they are so terrified of a class action suit in the wake of a data breach.

---

TRICIA HOWARD. Yeah, letting legislation or fear of retaliation run your security program, that sounds like a great move.

---

GRAHAM CLULEY. Yeah, it really doesn't, does it? But every hour of silence or hiding behind bland security advisories, the truth is that damages trust much more.

I think it's better just to say sorry because that hopefully will take the heat out of the situation.

---

TRICIA HOWARD. Agreed.

---

GRAHAM CLULEY. And poor customers are often left wondering if their data is being auctioned off to the highest bidder, all the while businesses are hiding behind their bland excuses which avoid saying sorry.

---

TRICIA HOWARD. Yep. And I think it's a huge misstep for a number of reasons.

One, first and foremost, your customers have a right to know where their data is going and who has accessed it. That's first and foremost.

Secondly, the security community really prefers that, and they respect that. We've seen that with— there was an outage on Discord, I think, a few years ago, and they were so good at updating where things were at.

They were super transparent. It was honestly a masterclass in how it should have been handled.

And I wish more companies did that because your point about bland, it's so— there's just zero substance whenever these things happen.

---

GRAHAM CLULEY. And it's impersonal.

---

TRICIA HOWARD. Yes.

---

GRAHAM CLULEY. I don't think people expect perfection. I mean, I don't think anyone sensible expects that there's such a thing as a hack-proof company, right?

You don't expect them to be completely impossible to breach. You know, in this day and age, being hacked is a fact of life.

It's more about how you respond to it afterwards. But people do expect, and customers and business partners do expect honesty and transparency, but so often they're let down.

And I think if someone came out and said, look, I'm really sorry, we fucked up, right? Here's what happened, here's what we're doing about it.

That almost always lands better with people than a sort of polished version of things. But that's what we keep on seeing.

That's why I was really pleased this week to see an organization actually say sorry after suffering a data breach. Checkout.com.

They are a payment processing service for businesses. They're just like a Stripe or a PayPal or a Klarna or one of those.

It allows firms to accept and manage money transactions regardless of whatever currencies or payment methods the customer wants to use. But being a big global payments processor like that does mean, of course, you've got a bit of a target on your back because you've got sensitive information, right?

When the cybercriminal's going, "Hehe, we like that." Money, money, money.

And sure enough, the hackers ended up stealing some data from Checkout.com. Now, it's thought that the criminals responsible were the notorious Shiny Hunters gang, and they apparently accessed data from one of Checkout.com's legacy third-party cloud file storage systems.

---

SNEHAL ANTANI. No.

---

GRAHAM CLULEY. And yeah, so Checkout.com, they estimate it affects less than 25% of their current merchant base, which could of course mean that they've lost a huge number of customers.

---

TRICIA HOWARD. No, I guess.

---

GRAHAM CLULEY. I don't know. But anyway, they say, look, of our current customers, only a quarter of you have to potentially be worried.

And they say it was mostly internal operational documents and other materials at that time. So not live payment data, which we're going through, not card numbers, thank goodness.

But it was an old system, and it was described as having been used up until about 2020. In other words, it was a system that should have been shut down long ago, but wasn't.

So they haven't used it for 5 years.

---

SNEHAL ANTANI. Wah, wah, wah.

---

GRAHAM CLULEY. Right. Sad trombone.

And according to CTO Mariano Albera, the hackers demanded this ransom to avoid the stolen data being leaked onto the darkweb. So this wasn't an attack where they encrypted files.

They simply stole the files. Some of these hackers now, they don't even bother with the encryption.

Why bother with the encryption? We'll just steal the data. We'll prove that we've got the data. We give people a sample back. We say, give us the money.

---

TRICIA HOWARD. Yeah, I mean, if that's the ultimate goal, yeah. Attackers are just like a business, right?

They have to find what is going to work for them. And if they're looking for the money, that's the way to do it.

---

GRAHAM CLULEY. But this is the thing which I really like. This CTO, Mariano Albera of Checkout.com, posted up on their blog, and you don't often expect to hear this from a major financial tech firm, but they didn't give any excuses.

They publicly disclosed the incident and they apologized.

---

SNEHAL ANTANI. Yes.

---

GRAHAM CLULEY. They actually used the word sorry. They said, we are sorry, full stop.

This was our mistake. We take full responsibility.

It's so refreshing.

---

TRICIA HOWARD. Yeah. I hope other companies or organisations take note of this because I was truly shocked whenever I saw this too.

---

SNEHAL ANTANI. Wow.

---

GRAHAM CLULEY. Can you imagine the response from their internal legal department? They'd be saying, "Oh, hang on. Whoa, whoa, whoa. We can't say that." There must be this internal battle going on saying, "Well, I think we should say it because we are sorry. We do accept that this should not have happened." I think it's brilliant.

But they went further than that. They said, "We will not be extorted by criminals. We will not pay this ransom," which again, I think is not a position which every company which has data stolen from it can take. I don't think it's always easy to make that decision. Yeah, but brilliant that they felt okay to say, "We own this, we're sorry about it, and by the way, we're not going to pay the ransom."

But what I really liked was they carried on and they said that they were going to donate the equivalent of the ransom which the cybercriminals had wanted to cybercriminal researchers at places like Oxford University and Carnegie Mellon. So they're donating all of that money to support their research into the fight against cybercrime.

So they've turned what could have been a PR disaster actually into a little bit of good news of them doing— I guess it's what corporations call CSR these days, isn't it? Corporate social responsibility. They're giving back to the community.

---

TRICIA HOWARD. Truly, way to go above and beyond. First, you say, "I messed up." Then you say, "Sorry about it." Then you say, "You know what? We're going to still give the money because we did mess up somewhere, but let's put it in a place where it could potentially have good in the future." That's phenomenal.

And what an incredible brand move, because if I were a customer, I would see that and say, "Wow, instead of saying we really care about security, they're literally putting their money where their mouth is." That's cool. That's awesome.

---

GRAHAM CLULEY. So maybe we should give them a little round of applause. Hang on. Hang on, Tricia. I don't think we should get too carried away.

Because when you dig into the details, there are a few uncomfortable truths. And the first thing that we can't ignore is that Checkout.com was using a legacy system. And legacy systems are a liability.

It's often not the new shiny platform that gets you hacked. It's the dusty bucket in the corner, literally, maybe a web bucket that no one remembers existed. It could be a file share you used to use 7 years ago but don't anymore. It could be a testing environment someone promised would be decommissioned next quarter but hasn't been. It could be a cloud folder from 2020 no one thought to lock down.

So although we're applauding them for saying sorry, and that's good, and we're applauding them for giving money to cybercriminal researchers and things, they did still lose data. They were a bit sloppy with their security, weren't they?

---

SNEHAL ANTANI. A little bit.

---

GRAHAM CLULEY. And cybercriminals, they love these forgotten systems because they're also often accessible and poorly monitored or unpatched and can be the route by which hackers can get into your organization and spread further and steal even more information. So we have to recognize that.

And transparency buys goodwill, but it doesn't erase the underlying security failure. To their credit, they communicated quickly and clearly. They took responsibility. They didn't try to apply some PR spin, but it was because this wasn't fully decommissioned.

So I like that they refused to pay the hackers. That's commendable. But you can't use that as your shield.

What I wouldn't like is any organization thinking, well, our stance of we won't negotiate with criminals is going to be a strong defense. Because one day cybercriminals might take something from you which is irreplaceable.

Yeah, it completely prevents your organization from running anymore, or it's going to lose you your customers, or you'll have to make staff redundant, or you cease operations, or your production line isn't working. So checkout.com actually got really lucky on this occasion because the data exposed was old. It was limited, but many companies won't be that fortunate.

---

TRICIA HOWARD. Yeah, true. And legacy systems are particularly tricky now with all the new technologies that we're introducing.

So many of the research pieces we do here are based on flaws that were known even some as 10 years ago. And they are being used now in new and different ways. And because of these things, you said, the dusty bucket in the corner, I love that.

Because of course, if I'm an attacker and my goal is to get the data, then to get money, I'm going to go in through the window if it's open rather than trying to knock down the locked front door.

---

GRAHAM CLULEY. Yep. I completely agree.

So what's my advice? I don't want to poo-poo too much. You know, they did say sorry, and I love that they're giving money to cybercriminals research, to the research into cybercrime.

It's the right way round, I hope, at least that money that they're giving. They're not funding a whole new era of cyber— Suppose we better just check the small print there.

But what can companies do to avoid this? Well, one of my pieces of advice would be to do a pre-mortem. Now, we're all familiar with the idea of post-mortems.

---

TRICIA HOWARD. Yeah.

---

GRAHAM CLULEY. You investigate something after a bad event. Why not do one in advance?

Why not ask yourself, what systems might we have forgotten about? What's still accessible that shouldn't be any longer? What would an attacker find easiest breach inside our organization, or what would embarrass us most if it were ever to leak?

Because most of the time, I think these attacks are grabbing the low-hanging fruit. So if you can secure that, you're going a long way to preventing yourself from becoming the next cybersecurity headline.

---

TRICIA HOWARD. Beautifully said. Beautifully said.

---

GRAHAM CLULEY. Thank you.

---

TRICIA HOWARD. It's almost like you have a show that you do for a living.

---

GRAHAM CLULEY. I wouldn't call it a living, but yeah.

---

TRICIA HOWARD. Okay.

---

GRAHAM CLULEY. Before we go any further, I need to share a quick word with you about one of our sponsors today, Vanta. You know how everyone's got an AI assistant these days? Well, imagine one that doesn't just write haikus about zero-day vulnerabilities, but actually does your audit work for you.

That is Vanta. It connects to all of your tools, gathers evidence, tracks compliance, and quietly helps you prove that yes, you do take security seriously.

Vanta automates all of that. It pulls everything together, keeps an eye on your systems, and basically makes sure you're ready for an audit at any time, which means no last-minute panic for screenshots and policies.

It also plugs into the tools you're already using and flags up issues before they become a right old mess. So if that sounds something that might save you from a few sleepless nights, check out vanta.com/smashing.

And if you use that link, you'll get $1,000 off. So don't forget, vanta.com/smashing.

And thanks to Vanta for sponsoring this week's episode. On with the show.

Tricia, what's your story for us this week?

---

TRICIA HOWARD. Well, mine is less uplifting, I would say.

---

SNEHAL ANTANI. Oh.

---

TRICIA HOWARD. However, since the holidays are upon us, at least over in this part of the world, travel is going to be up. And that means that, as we all know, this is whenever the attackers come out to play.

---

GRAHAM CLULEY. Yeah.

---

TRICIA HOWARD. So there is a new iteration in a series of malware incidents, if you will, that we've seen over the past few years that are targeting hotels and also hotel guests. So this is where it's not just I'm going after the hotel itself, or I'm not even just going after the booking platform itself.

I'm actually doing that as my first step and then moving on to attack the customer, the actual end user.

---

GRAHAM CLULEY. Yep.

---

TRICIA HOWARD. So this is actually a pretty sophisticated social engineering attack. How it worked is an attacker compromised an email account to send malicious messages internally through a booking platform to various hotels worldwide.

So this is really wild, right? It's not just geo-specific, which we've seen before.

This one is actually worldwide. Huge yikes.

---

GRAHAM CLULEY. These are all hotels using a central booking system. So one of the big names, which maybe we would use to book our own hotel rooms or vacations and things.

So the hotels are linked with them and they receive an email, which they believe to have come from that booking organization in some form.

---

TRICIA HOWARD. That's correct.

---

GRAHAM CLULEY. Okay.

---

TRICIA HOWARD. Yep. And it's sometimes to the point of editing the header to say that it was actually from that platform.

---

GRAHAM CLULEY. Okay.

---

TRICIA HOWARD. This is where the sophistication comes in because we are, I would say, largely aware as a society that we need to be looking at what we're clicking on. But it's really hard to do that whenever the things that you are clicking on look exactly the things you think you're supposed to be.

---

GRAHAM CLULEY. Yes.

---

TRICIA HOWARD. And so this was a really, really bad example of that. So the email was how it started.

It's a ClickFix campaign, but the ClickFix email contains a big bad URL. And upon being clicked on, kicks you to a page that downloads some malicious JavaScript.

And then that even checks for certain security measures so that it knows how to go about it. So if you do have an iframe, for instance, it will literally redirect you to a new page that does not have an iframe, does not have HTTPS.

Just goes over HTTP, and bingo bango, we land upon a super legitimate looking extranet. So this would be a partner portal or something for the hotels to go in looking at these booking platforms, right?

---

GRAHAM CLULEY. And that's where they grab the hotel's password or their login credentials or something that?

---

TRICIA HOWARD. Somewhat. So here's how the attack works. The person clicks on the email. It goes to a page that looks like Booking.com's extranet, even to the point of in the URL, using the words extranet or admin.

So this kind of suggests that the attackers probably did access the legitimate one so that they could mimic it, if nothing else. And then they do the reCAPTCHA tactic. Have you heard of this one?

---

GRAHAM CLULEY. Ah, this is—

---

TRICIA HOWARD. Yeah.

---

GRAHAM CLULEY. I love a CAPTCHA.

---

TRICIA HOWARD. I do too.

---

GRAHAM CLULEY. So I think we may have mentioned this on the show before. So the CAPTCHA is the thing which asks you, are you a robot or not? And it makes you jump through some hoops.

You maybe have to write some letters or press a key sequence or something. And with these particular ones, if I'm correct in saying, what they ask you to do is press a keyboard sequence.

---

TRICIA HOWARD. Yes.

---

GRAHAM CLULEY. Which copies a dangerous link into your clipboard and then runs some script which will take you to another page.

---

TRICIA HOWARD. It's a PowerShell command.

---

GRAHAM CLULEY. PowerShell command. Okay. So the ultimate end result of all this is you're going to have something malicious running on your computer.

---

TRICIA HOWARD. Yeah. So this is yet another step in this sophisticated social engineering thing. A good attacker knows that we created the internet and told people to click on things, and then we invented security and said, don't click on things.

That doesn't work. Then we add in new things to click on and the attackers mimic it too. Just the, I am not a robot.

Now this is wild because that in itself, for a typical user, clicking I am not a robot instills some sort of confidence. What it says to them is that, oh, this company cares about security. I mean, see how many times I can say that.

---

GRAHAM CLULEY. They take security seriously.

---

SNEHAL ANTANI. Yes.

---

TRICIA HOWARD. Very, very seriously. So they're actually not only abusing the platform, the hotels— I mean, there are so many people involved in this— they're also abusing a well-known, quote unquote, security tactic.

I mean, you don't get much worse than that. Upon doing the command, it downloads a zip file of a bunch of big baddie stuff, including a pure RAT malware.

It does a lot of stuff. It actually even sends status updates to the C2 so that the attacker can see what's happening in real time. And of course, sets up persistence, so it'll stay there for as long.

And it is fileless, so it's pretty difficult to detect. So the baddies know what they're doing because they even hosted the zip archive on a legitimate site that was compromised.

So they know on multiple steps here, they have created ways to have trust within the victim. And they're just luring them right in.

---

GRAHAM CLULEY. So the victim at this stage is still the hotel. So the hotel has been scammed.

---

TRICIA HOWARD. Yeah.

---

GRAHAM CLULEY. They've now got malware on their computers.

---

TRICIA HOWARD. Yes.

---

GRAHAM CLULEY. So how does it get to the hotel's customers?

---

TRICIA HOWARD. So this is part two. As the late Billy Mays would say, but wait, there's more. After getting over the actual hotels, they target the guests themselves using legitimate booking information that they got from the hotel that they had compromised. So they would reach out via email or WhatsApp and say something like, we saw your stay is on this date to this date. We actually, as an additional security measure, we need you to validate your card because people have been doing illegitimate things on bookings and we want to make sure you are secure.

So what that does, of course, we all know where this story goes. They click on the link that is absolutely malicious, and then the victim inputs their credit card information and other personal information into a phishing site, and that is then used to be compromised. Bingo, bingo, your attacker has all of your information.

---

GRAHAM CLULEY. And because the criminals have stolen the credentials of the hotel to log into the booking sites dashboard, they are able to send messages to hotel customers via that dashboard. So via their infrastructure. So any messages you get, which may be inside the booking app as well, really appear legitimate.

I know about this because I was out buying aubergines in the supermarket two years ago, and I got a message from Booking.com inside the Booking.com app. For a hotel trip, which I was just about to make, saying that they needed to revalidate my card. And thankfully, I was suspicious. But many people, because it would have arrived via the actual app rather than an email or a text, would have found that message really, really convincing.

---

TRICIA HOWARD. Oh, yeah. I mean, you're not only using the words and your typical social engineering tactics. Now, the method in which you're delivering the message is also a social engineering tactic.

As you said, it goes in through your app. We actually did a piece of research on that, specifically that one. There was a campaign that was going on as an extension of the original Info Stealer campaign that was targeting hotels through Booking.com sites.

They were saying that they were going to lose their reservation in the event that they did not revalidate their card. And they went so far, actually, as to have email correspondence directly with the victims, both at the hotel site and the actual end user victims, to create this trust. I mean, it was a very sophisticated campaign.

And I think what's particularly scary about this one is that, again, this is another evolution of something that has been in the works since — I think we did the research on it since 2023 was the first time it came out. And so many of the steps are actually as-a-service models. So now, not only are we having to defend against the people who are actually technically savvy enough to build the malware and/or edit whatever else is out there to fit their malicious purpose.

Now we're dealing with anybody who has an ax to grind and $40 on the darkweb. So it's a real problem. The malware itself was an as-a-service.

I believe there was a Telegram bot that was scraping credentials and stuff that was also a service. I mean, it effectively puts a little bow on cybercrime, doesn't it?

---

GRAHAM CLULEY. It really does. And I have to wonder whether these hotel companies are really doing enough because I've heard these complaints now about the way in which these companies are getting hacked and the way these messages are tricking travellers so convincingly for a few years now, you said yourself, and it feels it's still going on. The complaints are still coming in from travellers that they are being scammed in this way and they're not getting the best of support from the booking companies themselves.

In fact, there's been a complaint here in the UK that Booking.com have been replacing their customer service desk with AI chatbots.

---

TRICIA HOWARD. Yikes. I mean, it's just so tragic because especially, you know, we were talking about it around the holidays, which is like a good reason to travel, but there are also not super great reasons to travel, right? And if you're already in an emotional state, you are going to be more susceptible to this too, especially if it looks super legitimate and especially, especially if it's coming through your legitimate application that you booked it through. All of these things creates so much trust, and it's just scary.

---

GRAHAM CLULEY. You know what I'd really like, Tricia?

---

TRICIA HOWARD. What's up?

---

GRAHAM CLULEY. I'd like some of these hotel booking companies to say sorry.

---

TRICIA HOWARD. I was hoping you were going back for it.

---

GRAHAM CLULEY. Well, I'm delighted to be joined by another special guest this week, and our guest was the first CTO of JSOC. That's the part of the US military that runs special operations. His job was to build technology that helped real-world operators outthink real adversaries. And now he's bringing that mindset into everyday cybersecurity with Horizon3.ai. Snehal Antani, ex-CTO of JSOC, CEO of Horizon3.ai. That's quite the business card. Great to have you on the show.

---

SNEHAL ANTANI. Thank you for the opportunity. I appreciate it.

---

GRAHAM CLULEY. It's really fantastic to have you here. So I'm curious about this background of yours. How did you go from running military cyber operations to running a product demo for a Fortune 500 CIO?

---

SNEHAL ANTANI. I was actually in industry for almost my entire career. So I was at IBM, I was the CIO at GE Capital, and then I moved out west to be the CTO at Splunk. And my time at Joint Special Operations Command, in many ways, I was a tourist. You know, my military experience was watching Jack Ryan in Tropic Thunder. I didn't come from a military background. And so it was an absolute privilege and honor to serve within that community, recognizing that I had to earn a right to be in that organization every single day.

Now, keep in mind, within DOD, special operations and JSOC, they were already tech savvy. The bulk of their first mover of any sort of technology or capability or mindset is pioneered within the special operations community because that's how talented they are. And so the privilege of being able to help that organization get even better than they already were was amazing. I think the flip though is the influence and impact special operations experiences had on me as a leader and how much I grew personally and professionally from that role.

---

GRAHAM CLULEY. Yeah.

---

SNEHAL ANTANI. And how I now use that to try to be a better leader every day at Horizon3. And that's probably the bigger story out of this than anything else.

---

GRAHAM CLULEY. That's interesting. So now at Horizon3, you are doing some pretty interesting stuff. My understanding is it's all about autonomous pentesting. So let's say I'm a normal company. I do an annual pen test. I get a PDF report. I put it in my drawer and file it away. Are you suggesting that's not enough?

---

SNEHAL ANTANI. Yeah. You know, the goal of running a pen test isn't to find problems. The goal of running a pen test is to quickly fix problems that matter. And that's the fundamental mind shift that high-performing security organizations have over the laggard organizations that often end up in the news.

And so in cybersecurity, the only perspective that actually matters is the attacker's perspective, which is what the pen test brings to the table, right? And the attacker's perspective is how you're going to prioritize what to fix.

The attacker's perspective is how you're going to make sure your EDR or your SIEM or your WAF are actually tuned and working properly. In the attacker's perspective is how you're going to make sure your team has built the muscle memory to respond to a breach.

And that's the real epiphany for me in my time as a CIO, my time at Splunk, and then my time within the Department of War. And I think that the limiting factor throughout all that time was there are only about 25,000 certified ethical hackers globally.

There's only about 5,000 certified ethical hackers in the United States. Many of whom are in the military or in the government.

And so the amount of folks available to serve commercial organizations is very few. So that severe constraint of supply with an increased spike in demand makes it a very untenable situation.

And so you need some sort of force multiplier. And that was my bet. Could we invent some sort of AI system or autonomous system that could execute production pen testing of infrastructure at scale to be that force multiplier so humans can focus on the things humans are uniquely gifted at.

---

GRAHAM CLULEY. Right, so then you've got the resources, you've got the technology to be doing this effectively all the time. It's round the clock rather than a once a year approach.

---

SNEHAL ANTANI. That's exactly right. Our customers shift from 1 or 2 pen tests a year to 40 or 50 pen tests a month.

---

GRAHAM CLULEY. Wow.

---

SNEHAL ANTANI. And when I first saw that behavior, I was like, who on earth needs to run that many pen tests a month? And when you double-click, the journey is actually really interesting.

So the first thing is we all start with incomplete snapshots. We're testing a small slice of our network that's hopefully a reasonable sample of everything, and the problems we find there hopefully allow us to understand where else to look.

And so it was this incomplete snapshot, and the first thing people did with us was move to a comprehensive snapshot. Instead of running a pen test against 5 or 10% of their network, they can now run a pen test against their entire network as one large scope.

In fact, there was a large transportation authority in a big city in the United States that ran a pen test with 100,000 hosts in it. And these hosts represented subway stations, trams, buses, ticketing turnstiles, all of that transportation infrastructure.

So legit production environments. And they immediately found all sorts of exploitable problems that they had no clue existed in the network.

You know, we see dwell time data. And attackers, once they get in, they lurk. They take the time to map your entire network.

And in that mapping of the entire network, they become very precise in where to make the least amount of effort to cause the maximal amount of harm. And we don't do that from a pen testing standpoint, at least we didn't until we solved the comprehensive side.

Does that make sense?

---

GRAHAM CLULEY. Yeah, it makes a lot of sense to me. So you've got this autonomous pen testing platform. I think you call it NodeZero, right?

---

SNEHAL ANTANI. Correct.

---

GRAHAM CLULEY. And what does that do? How does it emulate the tactics and techniques that real threat actors are launching against an organization?

---

SNEHAL ANTANI. Yeah, it's a great question. So when you think about a cyberattack, on the one hand, we all are going to read news headlines about such and such a breach. But in that news headline or in that initial report, we're only told of the way the attacker gained initial access. Usually, right?

They gained initial access through this Palo Alto vulnerability that's now a Cisco KEV, or through this Avanti vulnerability or whatever else. That was just the way in. Once the attacker was in, they conducted reconnaissance and enumeration to discover everything that was network reachable.

They used techniques to harvest credentials, whether it was listening for credentials on the network. And they do these other steps to build basically a cyber terrain map of every host, port, service, defensive tool, credential, policy, and so on.

And it's from that map that they're able to effectively maneuver throughout the environment. And at the end, the outcomes or the impacts are fairly well defined.

The attacker's either going to become domain admin, which gives them the keys to the kingdom. They're going to find interesting sensitive data that they're going to pilfer and exfiltrate.

They're going to deeply borrow into some component to create the opportunity to attack at a future date of their time and choosing. And so it actually looks kind of like a chess game.

There are well-defined opening moves.

---

GRAHAM CLULEY. Yeah.

---

SNEHAL ANTANI. There are well-defined closing moves. In the middle of the chess game is completely dynamic.

And so as we built NodeZero, we thought of that in the same way. There are well-defined opening moves to a penetration test.

There are well-defined closing moves. But the middle needs to be completely dynamic based on what we've discovered in the environment.

And should we go after the router, the printer, or the television next? Well, that decision, that next best action, is based on discovered services, harvested credentials, historical record of success, likelihood of achieving our objective or our goal.

And so when you think about this as a graph analytics next best actions technical infrastructure, you can start to imagine what the underlying technology and system starts to look like.

---

GRAHAM CLULEY. So this isn't just about finding vulnerabilities. You're actually discovering attack chains here.

You're analyzing the movements which the hackers will commonly make and how they chain these things together.

---

SNEHAL ANTANI. Exactly right. More importantly, it's the combination of the attack path.

---

GRAHAM CLULEY. Yeah.

---

SNEHAL ANTANI. Attackers are going to combine a low plus a low to equal a critical, and that's how they're going to maneuver in the environment. It's about understanding the kill chain or the multiple steps the attacker took while also understanding the consequence of that exploitation.

Don't just tell me I've got a vulnerability on a host that doesn't tell me.

---

GRAHAM CLULEY. Right.

---

SNEHAL ANTANI. The attacker can chain together multiple issues that leads to the consequence of domain admin and consequences, how you're going to prioritize your resources.

---

GRAHAM CLULEY. Okay, so let's give some real-world examples here. You guys have run hundreds of thousands of autonomous pen tests for organizations.

There must have been somewhere everyone in the room goes pale. What was the most sort of, uh-oh moment which you've had with a customer where something has been unearthed?

---

SNEHAL ANTANI. Well, let's break this into probably a couple of subcategories because there's amazing stories.

---

GRAHAM CLULEY. Uh-huh.

---

SNEHAL ANTANI. So the first is the fastest, and this isn't about speed, but it is a testament to where the world is moving. Our AI hacker Node0, and I actually talked about this during my keynote at Black Hat, It got domain admin at a defense industrial base supplier in 77 seconds.

---

GRAHAM CLULEY. Oh boy.

---

SNEHAL ANTANI. No humans involved. Point, click, shoot, go. 77 seconds. That means if your defensive controls can't stop us in 76 seconds, it's game over. Because at second 77, we're going to get domain admin.

We're going to lock out all of your employees from the office. Gain access to all of your data, do whatever we want.

So 76 seconds. Think about that as a SOC.

Can you as a SOC analyst detect and stifle and contain an attacker in 76 seconds or less? The answer is probably not.

So we have an effectiveness problem in the future of cyber warfare is AI fighting AI with humans by exception. That's probably the craziest is the speed at which we're able to get there.

And I've got a few others. One was the coolest story.

---

GRAHAM CLULEY. Oh my goodness. Tell us.

---

SNEHAL ANTANI. Actually coolest in terms of we didn't expect this. There's this customer that had Windows Defender installed against 14,000 endpoints.

---

GRAHAM CLULEY. Okay.

---

SNEHAL ANTANI. And when they ran the pen test, of the 14,000 Defender agents, only one of them was misconfigured. Now we're going to be amazing job.

13,999 are correctly installed. We all did well.

Let's go on vacation. All it took was that one misconfigured agent for NodeZero to drop an implant, dump SAM, gain access to credentials, and then use that to laterally maneuver and eventually became domain admin.

One out of 14,000. And so when you think about your EDR and how crucial it is to ensuring your security, you have to make sure these EDR agents are actually configured and installed correctly.

And you can't trust that they're installed correctly. You've got to verify those configurations with the attacker's perspective.

---

GRAHAM CLULEY. You've got some pretty scary stories there, Snehal. Now, talking of scary things, and you've already touched upon this, the threat from AI, artificial intelligence, what's actually happening today in your opinion?

And how is AI changing both offensive and defensive security?

---

SNEHAL ANTANI. Let's start on the offensive side and bad guys. In September, Anthropic's team talked about a single person, not even a team, a single person creating various AI agents on Anthropic to execute a ransomware campaign.

And they were able to successfully ransomware 17 different organizations in this technique. And what this clever ransomware operator did was they built specialized agents in Claude that focus on one specific part of the attack, one for data discovery, one for evaluating the value of the data.

In fact, Claude recommended what the ransom price should be to the ransomware operator based on what they found in the data.

---

GRAHAM CLULEY. Oh my goodness.

---

SNEHAL ANTANI. Brilliant and amazing and terrifying. Now, on the one hand, of all the AI frontier labs, Anthropic is probably the most safety-focused, yet this operator was able to bypass that because in many ways, if you look at what they did, it was extreme compartmentalization.

What this ransomware operator did was create highly specialized agents that did one task, which on its own looks benign, but in aggregate across the agents is malicious. And so that's how they circumvented Anthropic safety mechanisms.

Now, if Anthropic makes their safety mechanisms more aggressive, then they're going to falsely identify good behavior. And what's the attacker going to do?

They're just going to further subdivide the task.

---

GRAHAM CLULEY. Yes.

---

SNEHAL ANTANI. To be even more specialized. In that example, Anthropic and the Frontier Labs enabled a single ransomware operator to operate like a large ransomware team.

So it was a force multiplier effect. I thought that was pretty remarkable when you think about the force multiplication effect an AI infrastructure can have for the bad guys, and the bad guys are embracing it faster than the good guys are.

---

GRAHAM CLULEY. This is the thing, isn't it? AI has not only democratized cybercrime, putting it in the hands of absolutely anybody, really. You don't have to be that much of a technical nerd anymore, but it's also multiplied the impact because things can be run on a bigger scale against more organizations. They don't need as much manpower as they used to need.

---

SNEHAL ANTANI. Yeah, exactly right. And that's a good segue into Villager. There's a great article on Villager that came out in September as well.

And Villager started as basically a capture the flag type tool that was written, I believe, by some folks in China. And it was an MCP server in front of Kali Linux.

And it allowed this capture the flag group or individual to quickly run various attacker commands in their CTF games. But that quickly expanded to be something like 4,200 different system prompts, plug-in architecture for different kinds of malicious things, whether it's keystroke logging and command and control and so on and so forth.

And then it was delivered through PyPI, the software distribution layer. And in September, there was something like 15,000 or 20,000 downloads of this already.

And this is a pretty legitimate standalone hacker tool that embeds DeepSeek as the AI model. And so this is an example where it dramatically lowers the barrier of entry into being a reasonably proficient offensive cyber person, especially if you're a bad guy.

So now if you combine these, you've got Villager as this AI-enabled hacking tool, you've got this Anthropic force multiplier effect, and you now have a massive increase of bad guy capability and capacity that the good guys are already struggling with the current state, let alone this new state that's emerging.

---

GRAHAM CLULEY. It sounds like bad news, but surely AI can also help us in our defense.

---

SNEHAL ANTANI. Yeah, for sure. So that's one part of it. There's a lot of hype around using AI for SIEM, for detection and response.

But I've actually not seen any impactful or useful application of that that actually works in real-world systems. At best, what I see is using LLMs to help process tickets better if you're in the SOC, but nothing that has truly transformed the reaction time of a defender yet.

We'll get there, but I think there's a lot of hype. And I think marketing is way beyond what actuality is.

Where I have seen some pretty incredible capability though, is the use of AI to improve source code security. And if you look at Aardvark from OpenAI that just came out, Claude Code has been doing this really well for a while.

It's completely transforming the way you do static application security testing, being able to find bugs in software, and hopefully soon being able to quickly triage the bugs in software, which is actually the long pole in the tent when it comes to secure coding. So I think that's where we're going to see the biggest effect, but that only helps improve current code being deployed.

But what about all the legacy code in firmware and other infrastructure that doesn't get patched frequently enough?

---

GRAHAM CLULEY. Well, it's good to hear that future code may be better written than it has been historically. Finally, you guys are basically sitting on a giant map of how companies are actually getting hacked, not theoretical vulnerabilities, but real-world weak points.

Imagine I'm a business leader listening to this podcast on the train. What should I be asking my security team on Monday morning that will tell me if we are actually secure?

---

SNEHAL ANTANI. It's a great question. So the commander I had the privilege of serving under in special operations used to say to me, don't tell me we're secure, show me, and then show me again tomorrow, and then show me again next week because our environment's always changing and the adversary always has a vote. The first thing every CIO and CISO should do is repeat those words to their team.

Don't tell me we're secure in PowerPoint or in some arbitrary stats or some compliance checkbox. Show me we're secure. Use the attacker's perspective to make clear the consequences of exploitation and the risk we are accepting as a business, or that we need to go surge dollars to, to make sure we remediate or mitigate. That's number one.

And then number two is how quickly are you fixing security problems? How often are they reoccurring and why? Proactive cybersecurity is all about staying in shape. You know, there's this quote, it's easier to stay in shape than to get in shape.

You need to build this cadence as a CIO or a CISO. Every single Monday morning, you should be looking at pen test results across your environment. You should be understanding how many problems you have, how quickly you're fixing them.

---

GRAHAM CLULEY. Great stuff. Well, listeners, you heard all about it here. And if you want to see what we've been talking about in action, Horizon3.ai are offering listeners a free autonomous pen test demo of Node Zero.

It runs the same attacker-style tests that we discussed continuously, not just once a year. It's a simple way to find the issues that actually matter before someone else does. So to grab hold of that and give it a try, go to Horizon3.ai.

And all that remains is for me to thank you for joining us on the show today, Snehal.

---

SNEHAL ANTANI. Amazing. Thank you for the invitation.

---

GRAHAM CLULEY. And welcome back. And you join us at our favourite part of the show, the part of the show that we like to call Pick of the Week.

Pick of the Week is the part of the show where everyone chooses something they like. Could be a funny story, a book that they've read, a TV show, a movie, a record, a podcast, a website. Or an app, whatever they wish. It doesn't have to be security-related necessarily.

Well, my pick of the week this week is not security-related. Regular listeners will be well aware, probably all too familiar with the fact that I'm a huge fan of just two things in the world, which are the game of chess and classic Doctor Who.

And in particular, when it comes to Doctor Who, I'm a big fan of the 1960s episodes made in black and white. When William Hartnell and Patrick Troughton were steering the TARDIS through the galaxy. But what you may not be aware of is that the BBC, in their great wisdom, junked a whole load of episodes way back when.

And many of them are still missing, although some have been found over the years, including in the basement of a Mormon church.

---

TRICIA HOWARD. Wow, amazing.

---

GRAHAM CLULEY. And inside dusty TV archives in Nigeria. 97 of the episodes are still currently missing. It does sound like there could be news around the corner that maybe we're going to find another couple of episodes soon, but 97 episodes are currently missing.

Now, fortunately, we have the audio of every classic Doctor Who story, but not the moving pictures. But what I can tell you is that there is a chap, his name is Philip Boyes, and he has devoted the last 3 years of his life to animating one of those missing episodes.

It is the first episode of the classic Doctor Who story, first broadcast by the BBC in 1965, called The Daleks' Master Plan.

---

TRICIA HOWARD. Exterminate.

---

GRAHAM CLULEY. Indeed. Uncanny. And he did this, and I love this, he did it without using any artificial intelligence.

---

TRICIA HOWARD. Snaps.

---

GRAHAM CLULEY. I mean, the truth is, after three years, he's probably thinking, I wish I had used artificial intelligence. Fair enough, I could have done it in three days instead. So it's a true labour of love.

It is up on YouTube. It costs you nothing.

He hasn't even monetised the video because he'd probably get in trouble with the BBC if he tried to. I mean, it's not Pixar quality animation, but by heck, I think it's extraordinary what he's done.

When the people are involved, it does look a little bit Thunderbirds. It looks a bit Supermarionation.

But when it's the Daleks and when it's the jungle and when it's spaceships, it's incredible. And I think it's great that there are people dedicated enough to their hobby to take restoring lost episodes with their bare hands and producing this.

It's a work of art. No financial incentive, I think it's marvellous.

So well done to Flip Boys and his reimagination of The Daleks' Master Plan, Episode 1: The Nightmare Begins. And I will link to it in the show notes, and that is my pick of the week.

---

TRICIA HOWARD. I'm honestly very excited to watch that. It's going to be so good.

---

GRAHAM CLULEY. Oh, you're a Doctor Who fan too? You like a bit of Who?

---

TRICIA HOWARD. Boo-wee-oo! Sure am.

---

GRAHAM CLULEY. Excellent.

---

TRICIA HOWARD. I will say I came in on the reboot.

---

GRAHAM CLULEY. Yeah.

---

TRICIA HOWARD. My first Doctor was David Tennant, but I really started watching it at the Christopher Eccleston episodes.

---

GRAHAM CLULEY. Yes.

---

TRICIA HOWARD. I haven't watched many of the older stuff, but man, so good. Such a good show.

---

GRAHAM CLULEY. Well, there's a lot to plough through of the older stuff, and it is a fair bit slower. So if you're of a certain vintage, you prefer it dawdling along a bit more.

Twenty-five minutes per episode. Anyway, go and check it out. Tricia, what is your pick of the week?

---

TRICIA HOWARD. My pick of the week is also not security related.

---

GRAHAM CLULEY. Good.

---

TRICIA HOWARD. It is actually Being Eddie. It is the Netflix documentary that just came out on Eddie Murphy.

---

GRAHAM CLULEY. Okay.

---

TRICIA HOWARD. I have been an Eddie Murphy fan since childhood. I mean, Dr. Dolittle was and still is one of my favorite movies.

And I mean, what an icon, right? I came in at a different part of Eddie's career than his standup.

---

GRAHAM CLULEY. Yeah. Which wasn't so child-friendly as I remember.

---

SNEHAL ANTANI. Yeah.

---

GRAHAM CLULEY. When you said you became a fan of him as a kid, I was thinking, hello? But okay, yeah, Dr. Dolittle.

That's right. Or that donkey which he played in Shrek, of course.

---

TRICIA HOWARD. Yes, of course. Of course, of course, Shrek, also one of my favorite films.

Yeah, he's just super iconic. So when I was growing up, he was all over the movies and the family-friendly movies, because this was after he had his kids, and so that became his focus.

And the documentary I think is really, really nice. First off, I mean, it's star-studded.

It's such a testament to how much Eddie has shaped the industry and shaped comedy in general and especially for people of color. It's really, really fascinating.

Dave Chappelle speaks on it. Jerry Seinfeld's on it as well, and they all talk about how Eddie just really changed the game.

And it's really heartwarming to hear from him as well, how he stayed on the straight and narrow and how he was able to avoid some of the truly tragic means of some people that reached his level of fame as well. And that was how he was able to manage it.

It's honestly just very heartwarming, and he is brilliant. I mean, getting to hear from him directly is such a gift anyway.

Yeah, it's certainly worth a watch.

---

GRAHAM CLULEY. Oh, okay. So it's called Being Eddie, and it's up on Netflix?

---

TRICIA HOWARD. That's correct. Yep.

---

GRAHAM CLULEY. Fantastic. Well, that just about wraps up the show for this week. Thank you so much, Tricia, for joining us. I'm sure lots of our listeners would love to find out what you're up to and follow you online. What's the best way for them to do that?

---

TRICIA HOWARD. Well, thank you for having me. This is so exciting. The best way to find me is on LinkedIn, actually. So LinkedIn slash Tricia Kicks SaaS is how you will find me. That's where anything that I'm doing, work or otherwise, will be.

---

GRAHAM CLULEY. Super duper. And you can find me, Graham Cluley, on LinkedIn as well, or follow Smashing Security on Blue Sky. And don't forget to ensure you never miss another episode, follow Smashing Security in your favorite podcast app, such as Apple Podcasts, Spotify, and Pocket Casts.

The episode show notes, sponsorship info, guest lists, and the entire back catalog of more than 443 episodes. Check out smashingsecurity.com. Until next time, cheerio, bye-bye.

---

TRICIA HOWARD. Cybernara.

---

GRAHAM CLULEY. You've been listening to Smashing Security with me, Graham Cluley. And I'm really grateful to Tricia Howard and Snehal Antani for joining us this week.

And this episode sponsors Vanta and Horizon 3 AI. And of course, to all the chums who've signed up for the Smashing Security Plus over on Patreon for their support of the show.

They include Michael Crumb, Tim de Rooyck, Skidone, Sital, Bobby Hendrix, Ashley Woodall, Marvin71, Govind Acharia, Christo V, Frankie Guzikowski, Funky Duck, Dan H, MJ Lee, and Richard Van Liesum. Well, what a marvelous bunch of people they are.

And if you would like to hear your name read out occasionally at the end of the show, all you've got to do is sign up for Smashing Security Plus. You will get early access to episodes without the ads.

All you got to do, just head over to smashingsecurity.com/plus for all of the details. Now, you can support the show in other ways.

For instance, you can tell your friends all about Smashing Security and encourage them to subscribe. Or you can leave us a 5-star review.

We had a pretty bad one the other day, and every time I go to look at our reviews, it stings my eyes, everybody. So if you were to leave us a nice review, then maybe you would fill this sad colander of a heart of mine.

You know, you pour in the love, it drips out the bottom, but you could plug one of the holes in my colander by leaving a lovely 5-star review on somewhere like Apple Podcasts. Anyway, no pressure.

Don't feel like I'm begging you too much. Every little bit helps.

I will be back next week with another episode, and I hope you will be joining me. Until then, cheerio, bye-bye.

-- TRANSCRIPT ENDS --