Canva - The Feedback Tool With a Key to the CRM, and the Regulator That Went First
17 minAudio
Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! A catch-up episode, recorded after the fact; developments after 23 September are flagged. Canva's own platform wasn't breached, but enterprise customer contact details and contract documents were exposed through Canny, a customer-feedback tool Canva had connected to its Salesforce CRM. Timeline. 28 August: Canny tells another customer, VRChat, that an unauthorised party accessed one of its internal systems; VRChat says forensics found no further activity after that date. 29 August: Canny tells Canva it is investigating unauthorised access; Canva removes Canny's Salesforce access immediately. 17 September: Türkiye's data protection authority publishes a notice (board decision dated 16 September) on Canva Pty Ltd's breach notification, saying 424 organisations in Türkiye are affected and the number of individuals is not yet determined; Turkish media report it, with a Canva statement. 21 September: Capital Brief reports the breach. Data involved. Names, business email addresses, workplace locations and work phone numbers of enterprise customers' staff; and, where shared, order forms, data protection agreements, master service agreements, invoices and business correspondence. Canva says its accounts, passwords, designs and content were not affected. How Canny was breached has not been disclosed; no CVE is involved. Later (from 23 Sep US time), unverified. DataBreaches.net reported a group calling itself The Seven Deadly Sins listed Canva on a leak site, and in a 26 Sep update said the group had supplied about 3 GB of alleged data, not independently validated. The group's claims of 2M+ Salesforce records and 200M+ warehouse rows are unverified. Whether Canva notified the OAIC is not public. Lessons. Review every app connected to your CRM and what it can read. Be able to revoke a vendor's access in minutes. Business contact details are personal information under the Privacy Act, and leaked contracts and invoices set up invoice fraud - warn customers early. If you operate internationally, a foreign regulator may publish first. Visit www.kinsoft.com.au to talk through your security and IT needs. Sources: Türkiye Personal Data Protection Authority; Türkiye Today; Webtekno; VRChat; Capital Brief; BeyondMachines; DataBreaches.net.
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.
So, I mean, whenever a really major corporate data breach hits the headlines, you know, we kind of have this tendency to visualize it like a blockbuster bank heist. Oh, yeah, totally. Like lasers and vaults. Right, exactly. You picture this sophisticated crew of threat actors battering down the digital front door, you know, bypassing biometric scanners and just making off with the gold bullion while all the alarms are playing. Yeah, which is a great movie plot, but... It really assumes this centralized, fortified perimeter,
like this binary state where the wall is either intact, keeping everyone out, or it's breached. But, I mean, stepping into the actual reality of a modern enterprise tech stack today, that whole bank vault analogy just completely falls apart. We aren't really dealing with single fortresses anymore. No, not at all. You're looking at a digital supply chain that is, by design, fundamentally porous. Yeah. I mean, the modern enterprise doesn't really have a single perimeter. It has thousands of micro perimeters. They're all woven together
by, you know, API connections and OAuth tokens and third party integrations. The walls aren't walls. They're like a membrane. Exactly. A very permeable membrane. And so that brings us to the core of today's deep dive. We're looking at this fascinating cybersecurity event from September 2026 involving Canva. Which is just a massive global design juggernaut. Right. And we're pulling from some really stellar technical analysis out of Kinsoft that breaks down how they found themselves in the middle of this massive
data exposure. But the twist here, the part that really forces you to reevaluate your own systems, is that Canva's actual platform was never breached. Yeah, that's the wild part. The attackers didn't touch a single pixel of Canva's core design data. Nothing. They didn't compromise the user accounts, no passwords, no identity access management, none of that internal infrastructure. So for everyone listening who, you know, manages their own SaaS sprawl, the big question is, how does one of the world's most recognizable software
companies get their sensitive data stolen without a single hacker touching their actual platform? And to understand that, you really have to look away from the main fortress. You have to look at the integrations. Right, the side doors. So Canva uses Salesforce as its CRM, which makes sense. Yeah, standard architecture. You use a heavyweight CRM to hold your major enterprise contracts, order forms, contacts, all that administrative lifecycle stuff. Okay, but where the vulnerability actually came in was through this third -party customer feedback tool called Kani. Canva connected
Kani to their Salesforce environment. And, I mean, the operational logic makes total sense, right? You want a feature request in Kani to automatically map to the specific enterprise customer in Salesforce. Sure. You want the tools to talk to each other. But this feels a bit like, I don't know, hiring a dog walker and accidentally giving them the master keys to your safe. Why on earth does a customer feedback widget have the permissions to read sensitive master service agreements and invoices in the first place? Well, what's fascinating here is that this touches on probably one of the most persistent structural
flaws in cloud integrations today. Which is? The over -provisioning of OAuth scopes. Oh, right. Yeah. So when an IT admin or like a product manager goes to set up a connection between two SaaS products, they're usually just presented with a set of default permission requests. The API needs authorization to read and write to the CRM to actually work. And let me guess, those default permissions are basically never granular? Seldom. I mean... SaaS vendors want their integrations to be totally frictionless. If they restrict
the API scopes too tightly, things break. And then you get angry support tickets. Exactly. It frustrates the user. So they just request this really broad read and write access. And the Canva admin clicks authorize. And suddenly, this lightweight feedback tool has a persistent, authenticated token. Wow. So it's just got a backstage pass to the CRM's deepest administrative databases. Pretty much. The principle of least privilege is just completely sacrificed for operational speed. You just want it to work. Which perfectly
sets the stage for the actual compromise. So let's trace the timeline here, because I think the speed of this is what really redefines modern incident response. It's incredibly fast. Yeah. So on August 28th, a completely different Canny customer, the social platform VRChat, they get notified by Canny about some unauthorized access to Canny's internal systems. Right. Then the next day, August 29th, CanE notifies Canva that they're investigating a breach. And Canva actually reacts immediately. They revoked CanE's access to their Salesforce environment in minutes. Yeah,
I mean, cutting that token authorization that quickly is textbook. It's the perfect incident response. Identifying the compromised pipeline and killing the session is exactly what their security operations center should be doing. But they didn't dodge the bullet, did they? Not at all. Because in the physical world, if you catch a guy breaking into your office and kick him out three minutes later, he can only physically carry so many files out the door. Right. He's dropping folders on the way out. But an API operates more like a digital printing press hooked up to a fire hose. It's a matter of machine speed
exfiltration. We aren't dealing with a human manually clicking through folders here. Right. An attacker who has compromised Kani's internal infrastructure can leverage that trusted API token to just... script automated database queries against canvas salesforce because to salesforce it doesn't even look like an attack does it exactly to salesforce it just looks like canny doing its normal job the api requests are fully authenticated they bypass all the external firewalls and intrusion detection systems because that trust was explicitly
granted man So a scripted query can just blast through thousands of records in, what, milliseconds? Milliseconds. Parsing and extracting names, emails, attached documents. By the time Canva got the alert on the 29th and revoked the token, the data transfer was already completely done. So they slammed the door shut, but the room was already emptied out. Basically, yeah. Okay, so the connection gets severed, the bleeding stops, and then silence. Yep, total radio silence. For weeks. The public hears... Absolutely nothing.
Which brings us to what I think is the most fascinating geopolitical angle of this whole thing. Oh, the regulatory nightmare. Yes. Because if you're an Australian tech giant dealing with a data exposure, you kind of assume you control the timeline, right? You prep the legal team, you draft the holding statements, you privately inform your big clients. You manage the narrative. Right. But instead, the entire story gets broken on September 16th by a government regulator in Ankara. The KVKK. Yeah. The KVKK just publishes a public notice detailing the whole breach notification
they got from Canva. They even outlined that 424 Turkish organizations were affected. Which is wild. It is. So a Turkish media outlet runs a story on the 17th. And meanwhile, Australian financial media didn't even catch wind of a massive homegrown company's breach until September 21st. Yeah, Canva essentially got scooped on its own crisis by a foreign regulator. Which has to just completely upend the traditional corporate communications playbook, right? It destroys it. Historically,
a company could contain the PR fallout while they, you know, spent 30 days doing a forensic investigation to find the exact blast radius. Just keeping it under wraps. Right. But if we connect this to the bigger picture, you just cannot do that anymore when you're operating under the fragmented realities of global data sovereignty laws. Because all these different regulators have totally different mandates on transparency. Exactly. The regulatory landscape is just a minefield of conflicting obligations now. Under frameworks like GDPR or the KVKK in
Turkey, the notification windows are incredibly tight, like 72 hours tight. Wow. And crucially, some of these bodies view their primary mandate as public protection, not your corporate confidentiality. So they will take your mandatory confidential disclosure and just publish it directly to their national registry to warn their citizens. So it basically renders the whole PR embargo thing completely obsolete. Totally obsolete. If you're a multinational, you just have to assume that the moment you file a compliance disclosure in
Europe or the Middle East, you've basically just published a global press release. That is such a huge takeaway for anyone listening. You lose the narrative control. Your crisis plan has to assume that local disclosures are going to leak globally before your own forensics team is even done. If your incident response plan relies on media silence while you investigate, your plane is just structurally broken. Okay, let's shift gears from the PR fallout to the actual payload, because we need to talk about what they walked away with before that API was severed. Right,
the data itself. So the Turkish regulator and Canva's own communications eventually clarify that the hackers grabbed staff names, business emails, workplace locations, and work phone numbers for employees at Canva's enterprise clients. Which is bad enough on its own. Yeah, but more critically, they exfiltrated order forms, data protection agreements, master service agreements, and historical invoices. And that, the exfiltration of those MSAs and historical invoices, that is a massive escalation in risk for those enterprise
partners. And we should mention, things got pretty theatrical in late September, didn't they? Oh yeah, the dark web claims. Right. So there's this threat actor group calling themselves the Seven Deadly Sins. They list Canva on an extortion leak site, and they're claiming to have three gigabytes of data, like two million Salesforce records and 200 million rows from a data warehouse. Yeah, but we have to put a massive caveat on that based on the source analysis. Right, because these claims are completely unverified, aren't they? Totally unverified. There is no CVE vulnerability score tying them to a specific technical exploit
on Candy's end yet. And honestly? Those numbers vastly exceed the scope of the API exposure that Canva has actually acknowledged. So why do they make these huge claims? Just for attention? It's extortion marketing. Threat actors on these leak sites frequently inflate their inventory. It's totally psychological. By claiming 200 million rows, they're trying to force a panic -induced settlement from the company. Or they just use those inflated metrics to build credibility with other cybercriminals. So basically, until forensic
telemetry actually validates it, treat it with a grain of salt? Treat it as marketing. Okay. So putting the unverified 200 million rows aside, let's look at the verified payload. They didn't get passwords. They didn't breach the graphic designs. Right. They just got business contact details, MSAs, and some invoices. And I think there's this common assumption that, you know, if the intellectual property and the passwords are safe, it's kind of a low -impact breach. Yeah. A lot of people think that. But you're saying an exposed MSA is actually more dangerous
than a hashed password. Without a doubt. This raises a really important question about how we value data. A compromised password can just be neutralized. You force a mandatory reset, you enforce multi -factor authentication, and it's useless. True. But a compromised master service agreement, that provides an adversary with the exact architectural blueprints to bypass technological defenses entirely and instead exploit human trust. We're talking about business email compromise here, right, BEC? Exactly. The mechanics
of modern BEC rely entirely on weaponized context. Weaponized context. I like that phrase. Think about a traditional phishing email. It's generic, usually poorly formatted, and it gets flagged by your email security gateway instantly. Or just by the human eye. Yeah, the classic, dear sir, please click link stuff. Right. But now, place that stolen Canva MSA into the hands of a sophisticated syndicate. They aren't going to send a generic payment link. They're going to engineer a scenario that seamlessly blends
into the victim's routine financial operations. Exactly. They target the accounts payable department of the enterprise client. And they draft an email that flawlessly mimics the tone and formatting of a routine Canva communication. Because they literally have the old emails to copy from. Exactly. dictates the payment terms. They attach a forged invoice that reflects the precise billing cycle, the correct internal vendor ID, and the exact dollar amount the client is already expecting to be billed that quarter. Wow. And the only
altered variable is the routing number at the bottom of the invoice. Yep. Just quietly redirecting that six -figure wire transfer to an offshore account controlled by the syndicate. And the account's payable clerk just authorizes it because it matches every single internal validation metric they have on file. Exactly. The security friction is completely bypassed because the data they used to build the fraud was entirely legitimate. And this is why B2B administrative data is so incredibly valuable on the criminal underground. It enables scalable, high yield financial fraud.
It really does. It transforms a technical API vulnerability into a direct financial liability for every single enterprise connected to that CRM. The blast radius extends way, way beyond Canva's own balance sheet. So Canva now has to aggressively educate its entire client base on invoice redirection fraud because the bad guys have the exact documents needed to pull it off perfectly. They do. It forces everyone into a proactive defensive posture. Man. So tying all
these threads together, I mean, this Canva incident just provides such a visceral look at the fragility of our modern infrastructure. It really forces anyone listening to evaluate their own digital perimeter or, I guess, their lack of one. Yeah, there is no perimeter. We've established that every connected app, every little third -party widget, every authorized API token, they're all load -bearing pillars in your security architecture now. The ecosystem is literally only as resilient as the least secure vendor you've granted access to. Which is terrifying when you think about
it. It is. So for you listening, you know, you have to audit those integration scopes. You have to have the operational agility to kill a compromise session in minutes, knowing that machine speed exfiltration has probably already happened. Yeah, speed is everything. And you have to build these crisis communication frameworks that actually anticipate foreign regulators publishing your breach before your own PR team has even drafted the memo. Right. You don't control the global PR order anymore. Exactly. And above all, you really have to recognize that administrative
business data, the contracts, the invoices. That is the primary ammunition for the most devastating financial exploits today. But, you know, I'd argue the challenge goes even deeper than just auditing the integrations that your IT department. actually officially sanctions. Oh, how so? Well, think about it. The Canva and Cani connection was an approved documented integration. It was mapped out and known to their architecture team. Which brings us to the unknown ones. Exactly. The shadow integrations. Every organization right
now is grappling with shadow IT. It's just the proliferation of tools that employees adopt completely outside the purview of the security department. Just consider your own workflow or anyone listening. How frequently do you or your coworkers authorize a third -party plug -in to, I don't know, synthesize meeting notes or convert file formats? Oh, man, all the time. You just click Allow Access to Google Workspace or Connect to Microsoft 365 because you just need the stupid thing to work for a 2 p .m. deadline, right? You completely
bypass the procurement review. We all do it. But... Every single time that authorization is granted, a new API token is generated, usually with really broad read and write scopes, just connecting this totally unvetted external server directly into your core identity fabric. Wow. So the next major supply chain breach that dominates the news cycle, it might not originate from an enterprise -grade vendor like Canny. Right. It is highly probable it's going to stem from a
free unpatched browser extension or a productivity widget that... Some employee just plugged into their corporate environment yesterday afternoon, bypassing every security control you thought you had. That is, wow. That's a really sobering reminder that in this hyper -connected ecosystem, convenience is really just vulnerability by another name. It truly is. Well, thank you for joining us for this deep dive into the complexities of the digital supply chain. It's been eye -opening. Yeah, thanks for having the discussion. For everyone listening, take a moment today to actually review
the third -party apps authorized on your own accounts. Audit those permission scopes and really question whether that random PDF converter actually needs unrestricted access to your entire email history. Stay curious, stay vigilant, and we will catch you next time.
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...
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...
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...
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...
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;...