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

Tech Talks With Kinsoft

Tech Talks With Kinsoft artwork

PaperCut - 9:42 on a Thursday, Three Emergency Patches, and the Bug the Founder Wrote Himself

Sep 8, 202622 minAudio

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! A Melbourne software company, a print server, and three emergency patches in six days. The company. PaperCut Software International, founded 1998 by Chris Dance and Matt Doran, headquartered in Melbourne with offices in Portland and London. Privately held, still founder-led. PaperCut's own figure: around 100 million users across more than 70,000 organisations. The timeline. 27 August, 9:42am AEST - an education-sector customer reports a compromised PaperCut MF server; a second organisation reports similar activity at 5:05pm, and its logs allow the full chain to be reconstructed. PaperCut declares a P0 and publishes an advisory deliberately stripped of technical detail. 28 August, 2:10am - Emergency Patch Release 1 for versions 25 and 26; version 24 later that day; CVEs assigned; Emergency Patch Release 2 the same day after watchTowr and Huntress bypass the first fix. 29 August - Defused honeypots see the attacks pivot to database-table dumping. 31 August - both CVEs added to the CISA Known Exploited Vulnerabilities catalogue with a 14 September federal deadline; a Metasploit module is published. 1 September - Emergency Patch Release 3, fixing two regressions and adding hardening. Install it even if you patched earlier. The vulnerabilities. CVE-2026-81578, authentication bypass in the web management interface (CWE-306), 8.8 High under CVSS v4.0. CVE-2026-82078, unsafe dynamic class loading in the database connector (CWE-470), 9.4 Critical under CVSS v4.0. Chained, they give unauthenticated remote code execution: an Apache Tapestry routing quirk lets an attacker display a public page while invoking an admin component, rewrite the external user-lookup settings, and abuse the bundled Apache Derby driver to reach an attacker-controlled H2 database whose initialisation parameter creates a JavaScript-backed trigger that spawns a process. What is not known. No attribution by anyone - not PaperCut, not Huntress, not watchTowr, not CISA. No ransomware gang has claimed it. No victim count, record count or data-category list has been published. Shadowserver tracked 800+ internet-exposed servers as at 1 September; that is an exposure count, not a compromise count. What to do. Patch to Emergency Patch Release 3. Assume compromise if the server was internet-facing and unpatched after 26 August - patching does not evict an existing intruder. Hunt PaperCut's indicators: suspicious activity from pc-app.exe, missing or truncated server.log, and the "no suitable driver found for jdbc:no:x" and "Database error looking up cardID" errors. There are no published network indicators. Then ask why a print server was internet-facing at all. Also timely: the exposure draft of the Privacy Amendment (Personal Data Protection) Bill 2026, released for consultation around 1 September, proposes a hard 72-hour notification deadline to the Information Commissioner. Visit www.kinsoft.com.au to talk through your security and IT needs. Sources: PaperCut security bulletin and Chris Dance's incident blog; Rapid7; BleepingComputer; Cyber Daily; SecurityWeek; CISA Known Exploited Vulnerabilities catalogue; iTnews.

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.

You know, if you were to walk into like almost any medium to large office building in the country right now, past a reception desk, past the break room, you would eventually find it. The absolute least glamorous, most completely ignored piece of technology in the entire building. The print server. Yeah, it just sits there in the corner, humming away, smelling faintly of ozone and warm paper. Right. It's just this boring beige box. Or, I mean, maybe these days it's a virtual machine tucked away on a server rack somewhere, just

quietly counting pages so that somebody in finance can eventually bill the right department for toner. Exactly. Nobody really thinks about it. No, because it doesn't hold the central customer database. It doesn't run payroll. It is practically invisible. But if you're listening to this and you think your company's data is safe simply because you have, like, a massive firewall around your main servers? We're going to spend this deep dive proving why you might be wrong. Today, we are looking at how this completely mundane piece of technology became the epicenter of a

frantic global cybersecurity meltdown. It really is a remarkable story, especially when you realize the sheer scale of the target here. We are looking at a vulnerability in software made by Papercut. And if you haven't heard of them, they're this massive but very quiet Australian software success story. Yeah. Founded back in 1998. Right. Right. 1998. And they have around, I mean, something like 100 million users across more than 70 ,000 organizations globally. It's huge. So for this deep dive, we are pulling from a highly detailed

autopsy published by Tech Talks with Kinsoft. They broke down exactly what happened when a routine Thursday morning turned into like. a six -day patching frenzy for IT teams around the world. We're going to explore the ingenious mechanics of how this hack actually worked, look at a surprisingly candid response from the company's CEO, and ultimately figure out what this means for the new data privacy laws coming to your workplace. Okay, let's unpack this. The timeline is really the best place to start because the sheer speed of a modern cyber incident totally

redefined how we think about defensive IT. yeah i was looking at this timeline from the source material and the escalation is just It's what really stands out. So we're going back to Thursday, August 27th at exactly 9 .42 a .m. Australian Eastern Standard Time. A customer in the education sector reaches out to Papercut. Right. Just to report that their server has been compromised. Exactly. And at this point, it's just a single support ticket. But by midday, Papercut realizes this isn't a fluke. They declare a P0, which

is their absolute highest severity incident level and is a zero day. And I mean, a zero. day is basically the worst case scenario for a software vendor. It means hackers are actively exploiting a flaw in the wild and the vendor has had literally zero days to prepare a fix. Right. So by 5 .05 p .m. that same afternoon, a second organization in the same region reports malicious activity. But crucially, this second organization has complete enough server logs for Papercut to reconstruct the full attack chain. They finally see the invisible door, basically. Yeah, they see how the attackers

are walking through. And that same day. Papercut rushes out an urgent advisory to its customers. But the Kinsoft breakdown notes something highly specific about this advisory. Papercut published it without an official identification number, like a CVE, without a severity score, and with almost zero technical details about how the hack was happening. Which is pretty standard for a zero day, actually. But I have to push back on that strategy, though, because it feels like... I don't know. It feels like telling a neighborhood

there is a burglar on the loose, but intentionally refusing to describe what they look like or what tools they are using to pick the locks. Like, how are IT teams supposed to defend themselves if they don't know what to look for? I get that. I totally do. But the logic of incident response during an active zero -day event is, well... It's brutal, but necessary. When you have an active vulnerability being exploited and you do not yet have a patch ready to deploy, releasing technical details doesn't actually help defenders as much as we might hope. Really? Because wouldn't

they know what to block? Well, think about it. An IT admin can't rewrite the vendor's source code on a Thursday night. What releasing those details does do... unfortunately, is immediately arm copycat hackers. Ah, OK. So you're basically handing the lock picking blueprint to every other criminal on the Internet before you've had time to change the actual locks. Exactly. If Papercut had explained the exact mechanism of the flaw on Thursday, thousands of other opportunistic threat actors would have reverse engineered it

and started using it by Thursday night. The modern cyber warfare landscape relies. Like heavily on automation. Once bloat is in the water, scanners are spun up globally to find vulnerable servers. OK, that makes sense. But even with withholding the details, the scramble over the next few days is just exhausting to read. So by Friday at 2 .10 a .m., Papercut ships emergency patch release one. But independent security researchers at companies like Watchtower and Huntress, they start pulling the patch apart to see what it changed. Because that's what researchers do. They reverse engineer the fix. Right. And they

quickly find multiple ways to bypass the first fix, plus an additional authentication bypass that the first patch completely missed. Yeah, because a patch is often just a bandage over a really complex wound. Modern software systems are incredibly dense, and researchers quickly found where the bandage was leaking, so to speak. And while the defenders are frantically trying to apply more bandages, the attackers are shifting their strategy. Right. Because by Saturday, security researchers running honeypots. And those are essentially decoy servers designed to safely

observe hacker behavior. Yeah, exactly. So Diffused, one of the organizations running these honeypots, observed the attackers moving away from just trying to execute remote commands. They started dumping database tables directly. They pivoted straight to data theft. Which is a massive escalation. Yeah. And then Monday hits and a Metasploit module gets published. And for anyone who isn't a penetration tester, Metasploit is basically an off the shelf framework. It democratizes the exploit. That's the best way to put it. Once a vulnerability

is packaged into a Metasquoit module, it is handed out to the masses. You no longer need to be this, like, sophisticated engineer capable of writing custom code to break into a Papercut server. It's just plug and play at that point. Right. The exploit is loaded into a standard toolkit, meaning pretty much any low -skilled criminal with an internet connection can just point and click their way into a network. Which forces Papercut to act yet again. So on Tuesday, September 1st, they release Emergency Patch Release 3. This one fixes regressions, meaning bugs that

the first two patches accidentally introduced, and it adds more systemic hardening. And they bluntly tell every customer to install it immediately, even if they already patched over the weekend. Yeah, that is three emergency patches in six chaotic days. So how did the attackers bypass these frantic patches so easily in the first place? I mean, how is it so broken? Well, to understand that, we have to look at the anatomy of the flaw itself. The reason the patching was so difficult is that this wasn't just one broken line of code. It was two... Highly specific vulnerabilities chained together in a way that tricked the software's

entire architecture. Yeah. The source material breaks down these two vulnerabilities, and the scoring framework is actually fascinating. First, you have a flaw in the web management interface, which is an authentication bypass. It gets a severity score of 8 .8 out of 10. Right. Pretty high. Then you have a flaw in the database connectors, specifically something called unsafe dynamic class loading. That one gets a 9 .4 critical score. But the CVSS scoring framework notes that the 9 .4 database flaw requires high privileges to execute. Which is the catch. Yeah, I found

that contradictory. How can an unauthenticated hacker remotely execute code if the flaw specifically requires high administrative privileges? That paradox is solved by the chain itself. On its own, the database flaw is somewhat protected because you have to be an admin to reach it. But the authentication bypass is the key. The first flaw supplies the privilege required for the second flaw. Oh, I see. Yeah, you use the 8 .8 vulnerability to trick the system into thinking you have admin rights. And once you wear that administrative mask, you just make a beeline

for the database driver. Okay, let's trace this chain because the Kinsoft report keeps it remarkably grounded. So PaperCut's web interface is built on a third -party framework called Apache Tapestry. Tapestry handles how web pages load, and it has a feature where you can request to display one page but run a component belonging to a completely different page behind the scenes. And that routing capability is incredibly common in web frameworks. It's meant to make loading efficient. But PaperCut, well, they only checked if the user was allowed to see the main display page. They forgot to

check if the user was authorized to run the hidden component. wow yeah that single missing authorization check is the fatal flaw in this entire saga So an attacker asks the server to show them the public error page. And error pages have to be public, otherwise the system couldn't show a basic error message to someone trying to log in. Right, that part is totally normal. But while asking for that public page, the attacker uses Tapestry's routing feature to secretly invoke an administrative component. Suddenly, without even typing a password, they have the power to

rewrite core configuration settings. And they use that illicit access to rewrite the settings. settings for the database driver. So Papercut ships with a database driver called Apache Derby. The attacker points this driver away from the local database and tells it to connect to a malicious database URL that the attacker actually controls. I was wondering why a database driver would just accept a random URL like that. But that URL opens a connection to an H2 database. And here is where it gets incredibly convoluted, like a digital Rube Goldberg machine. That H2 connection supports

an initialization parameter. Yeah, because when a database connection starts, it... often needs to run setup scripts so h2 allows you to run an initialization script embedded directly in the connection url and it allows that script to trigger a javascript engine So the server executes the JavaScript, and the JavaScript contains instructions to spawn an operating system command line process. The attacker triggers this entire cascade with a forged user search on a public web page, and boom, they are running malicious code on the operating system. Exactly. And the

most crucial takeaway from this technical breakdown is that once you get past that first missing authorization check, Nothing else in that chain is a bug. Wait, none of it? Literally none of it. Capistry's component routing is a documented feature. Derby's driver accepting a remote URL is a feature. H2 using an initialization parameter to run JavaScript is a documented intended behavior. That is wild. Yeah, everything downstream from that initial entry point was just software doing exactly what it was designed and programmed to do. So let me get this straight. I walk into

a bank. And I asked the teller to use the public restroom. Anyone is allowed to use it. But because of how the plumbing was designed, flushing the toilet in the public restroom automatically opens the main vault in the basement. That's a great analogy, actually. So the toilet and the vault are both working flawlessly as designed. They just absolutely should not be connected to each other. Right. The architect who built the bank. didn't design the plumbing or the vault from scratch. They bought a prepackaged bathroom unit and a prepackaged vault door, and they just didn't

realize the wiring overlapped. Modern software is basically constructed out of these Lego blocks, these third -party frameworks and libraries. Developers stitch together Tapestry and Derby and H2 to build robust applications quickly. But when you stitch all those complex Legos together, you create unintended interactions. Exactly. The most devastating hacks today rarely involve breaking the code in the traditional brute force sense. They are about understanding the complex architecture of those Lego blocks better than

the people who built them. It is entirely about tricking the code into executing its own features in malicious ways. That realization has to be a nightmare for the developers, just watching the perfectly good features they build get weaponized against them, which brings a fascinating human element to this whole story. On September 1st, Papercut's co -founder and CEO, Chris Stantz, published a public blog post. The title alone sets a tone. It was called Behind the Scenes, What Happened After 942 AM on Thursday. And that

is so rare. Corporate incident reports are almost universally drafted by legal teams who are absolutely terrified of a lawsuit. They use passive voice. They say things like a vulnerability came to exist in our environment. Right. Like nobody actually did anything. The flaw just materialized out of thin air. Exactly. Or they rely. on the classic phrase a sophisticated threat actor, which implies the attack was so advanced that nobody could have possibly stopped it. But Chris Stantz explicitly refused to blame the third -party Apache libraries. He admitted in writing that the missing authentication check at the

front door came from code that he personally had contributed to over the years. It really is a masterclass in corporate accountability. I mean, in a landscape where companies reflexively hide behind PR spin, this 28 -year -old privately held company basically took it on the chin. It's entirely disarming. People want to be angry at a faceless corporation, but it's really hard to maintain that rage when the founder is standing right there saying, my bad, I messed up the Lego blocks. Yeah, defenders and security researchers genuinely respect technical honesty. That transparency

neutralized the PR crisis and built trust with the community. But, you know, in that same blog post, Dance also admitted a massive blind spot. The company doesn't actually know who is attacking them. Yeah, there has been zero official attribution for this attack. Papercut doesn't know. Government agencies like CISA don't know. Security firms don't know. No loud ransomware gangs have stepped forward to take credit. No nation states have been named. Now, the source material does note that Papercut had a serious vulnerability exploited by named criminal groups back in 2023. But it

explicitly warns against using that past history as evidence for this current attack. Right. We simply do not know who is behind this specific one. We lack names. But we do have behavioral clues, right? Jake Knott, the head of threat intelligence at Watchtower, he noted that as the attack progressed, they saw a shift from automated vulnerability scanning to actual hands -on keyboard human interaction. Yeah, meaning real operators were manually digging through these compromised servers. And there was a detail about their methods that really caught my eye. The attackers held their malicious payloads in

memory rather than writing them to the hard drive, and they keyed those payloads specifically to themselves. Which is very clever. Yeah, they made sure that rival hackers couldn't piggyback on their exploit and reuse the compromised host. By locking the door behind them, they secured exclusive access. But why go through all this effort just to sit quietly on a print server? I mean, if you aren't dropping ransomware immediately to demand a payout and you are actively locking out rivals, what is the economic endgame here? Well, this introduces the specialized economy of initial access brokers or IABs. Not every

hack is a loud smash and grab operation that locks up your files. Often the most dangerous threat actors are actually the quiet ones. Jake Knott described them as aggressive outcome type operators. Aggressive outcome type operators. OK, what does that mean? Their entire business model is to quietly find a way in, secure that access, lock out the competition and then just wait. Oh, so they act as digital lockpicks, basically doing the hard work of finding the unlocked window, and then they sell the map to someone else. That is exactly it. They sell that access on dark

web forums to ransomware gangs or corporate espionage groups. Access to a hospital network might sell for thousands of dollars, while a local school district goes for maybe a few hundred. The IAB wants exclusivity because if another hacker trashes the server with clumsy malware, the IAB loses their asset. Right. The product is ruined. Exactly. They protect the compromised server from other hackers just so they can sell it to the highest bidder later. The scale of that marketplace is just concerning. Shadow Server, an organization that monitors internet activity, they were tracking

between 800 to 1 ,000 plus internet -exposed paper cup servers during this window. And those aren't all guaranteed to be victims. Some might be honeypots or patched systems, but that is a massive surface area of potential entry points that initial access brokers could be sitting on just quietly holding the keys. Which naturally leads to the actionable phase of our source material. What do IT teams actually need? to do right now to protect their infrastructure. So Kinsoft lays out four highly actionable steps. First, patch to emergency patch release three. Do not stop

at the earlier patches because, as we discussed, fixing one thing introduced regressions that broke other things. Second, and this is a daunting one, assume compromise. The source explicitly warns that patching does not evict an attacker who was already inside. Right, because a patch fixes the broken window. It doesn't kick out the burglar who's already standing in your living room. If an attacker used the exploit to create a backdoor administrative user account, well, applying the patch doesn't delete that account. So if your server was exposed to the internet

at any point between August 26th and when you patched, you have to actively hunt through your system. You're looking for suspicious activity from a process called pc -app .exe or looking for server logs that have been truncated or deleted entirely to hide tracks. Yeah, the log deletion is a huge red flag. And third, stop putting print servers on the public internet. I mean, a print server has absolutely no business being exposed to the open web. It should only be reachable from a local office network or through a VPN. A simple firewall rule dropping public traffic

would have made this entire zero day a non -event. for the vast majority of victims. It's just staggering that so many organizations left him exposed. And the fourth step targets internal bureaucracy. IT teams have to fix their patching speed. If your company relies on a change advisory board that meets every two weeks to approve software updates, you are structurally incapable of surviving a modern zero day. You really are. You need an emergency path to deploy patches in hours, not weeks. And that speed isn't just a technical

recommendation anymore. It is rapidly becoming a legal requirement. This incident is basically colliding with impending legal shockwaves regarding data privacy. Yeah, this hack happened at a really pivotal time for Australian privacy law. The Attorney General Espertment recently released a draft of the Privacy Amendment, the Personal Data Protection Bill 2026. This bill proposes replacing the old kind of vague standard of notifying authorities as soon as practicable with a hard 72 hour deadline. You have 72 hours to notify

the information commissioner once you are aware of an eligible data breach. And we really have to consider what kind of data. we are talking about here. Print servers hold massive amounts of personally identifiable information. Right. They store usernames, employee card IDs, metadata about who printed what. And if print archiving is enabled, they hold the actual printed documents themselves. We're talking about financial reports, HR complaints and patient records. OK, but I have to challenge how this legislation functions in reality. We just established that sophisticated

attackers like initial access brokers will likely delete your server logs to hide their tracks. And the vendor, Papercut, is telling everyone to just assume compromise if they were exposed. So if my logs are wiped and I... just have a blanket warning from a vendor to assume the worst, how can my IT team possibly confidently report a breach within 72 hours? Doesn't the word aware in that draft legislation do an impossible amount of heavy lifting? If I can't prove they stole data, I'm never technically aware. I'm just suspicious. And you've hit on the exact stress test that

this paper cut incident provides for the new legislation. If this draft becomes law in its current form, it will force a monumental shift. in corporate infrastructure. You can no longer rely on the server itself to keep its own logs. Because the attacker just deletes them. Precisely. If an attacker wipes the local logs, you lose your visibility and your ability to comply with the law. To survive this regulatory environment, companies will be forced to maintain off -site immutable logging. They will have to stream their server logs in real time to completely separate locked down systems. That is a massive logistical

hurdle. I mean, it elevates this from an annoying IT problem where you have to update some software into a board level legal liability. Companies have to build entirely new systems just to ensure they have the evidence required to meet government reporting mandates. Yeah, the burden of the Privacy Act. sits with the customer running the server, not the vendor who wrote the software. The organizations that thrive in this new era will be the ones that have rehearsed exactly who makes the call to declare an incident and how to verify it the second that 72 -hour clock starts ticking. It

really has been an incredible journey unpacking this. We started with A mundane print server, you know, this boring beige box in the corner counting pages. And we've watched it become a battleground where perfectly functioning software features were chained together and weaponized against themselves. It sparked three emergency patches in a single week, prompted a really refreshing spin -free apology from a CEO who took personal responsibility and highlighted the shadow economy of initial access brokers. Ultimately, it serves as the perfect test case for how companies will handle aggressive new data privacy laws. It is

a perfect. storm of technical complexity and human oversight. And it leaves us with something quite unsettling to consider. If a 28 -year -old totally unglamorous print server holds enough sensitive data to trigger a global incident response and start a 72 -hour legal countdown, Well, what other invisible pieces of technology in your workplace are sitting completely exposed to the open Internet right now? What else is just humming away in a server rack, waiting for its own 9 .42 a .m. moment? That is a very necessary thought to end on. Thank you for joining us for this

deep dive. And remember to keep questioning the world around you. Catch you next time.

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. Iris Ohyama - A Suspicious File in the Cloud, 7.7 Million Customers at Risk, and an Online Store That Went Dark

    Nov 19, 202620 minAudio

    Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Japanese home goods company Iris Ohyama (online store operator Iris Plaza) found a suspicious file in an external...

    Credit the people on this episode

  2. Who Gives A Crap - A Marketing Supplier Breached, Six Days to Tell the Client, and Customer Data That Didn't Need to Be There

    Nov 17, 202619 minAudio

    Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Australian online retailer Who Gives A Crap told customers on 7 November 2026 that their information was accessed...

    Credit the people on this episode

  3. Last Week in Tech - An Outlook Zero-Day That Leaks Your Password Hash, Chrome's Eighth Exploited Bug, and Nvidia Without Jensen Huang

    Nov 15, 202621 minAudio

    Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Last week in tech, 7 to 14 November 2026. Microsoft's November Patch Tuesday (10 November) fixes well over a...

    Credit the people on this episode

  4. Chainalysis - From "Non-Sensitive" to "All Users": How a Breach Notice Changed in Five Days

    Nov 12, 202619 minAudio

    Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Chainalysis disclosed on 30 October 2026 that an external actor using "offensive agentic techniques" accessed...

    Credit the people on this episode

  5. NSW Health - A Dental Voucher Website, Four Gigabytes Taken, and the Providers Whose Details Were Exposed

    Nov 10, 202623 minAudio

    Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! NSW Health reported unauthorised access to its Oral Health Fee For Service Scheme website on 29 October 2026;...

    Credit the people on this episode