Archive for author: makoadmin

We received a timely phishing email pretending to come from Home Depot. It claimed we’d won a Gorilla Carts dump cart (that’s a sort of four-wheeled wheelbarrow for anyone unfamiliar)—and said it was just one click away.

It wasn’t.

The whole image in the email was clickable, and it hid plenty of surprises underneath.

Sender:

The sender email’s domain (yula[.]org) is related neither to Home Depot nor the recipient.

The yula[.]org domain belongs to a Los Angeles high school. The email address or server may be compromised. We have notified them of the incident.

Hidden characters:

Below the main image, we found a block filled with unnecessary Unicode whitespace and control characters (like =E2=80=8C, =C3=82), likely trying to obfuscate its actual content and evade spam filters. The use of zero-width and control Unicode characters is designed to break up strings to confound automated phishing or spam filters, while being invisible to human readers.

Reusing legitimate content:

Below the image we found an order confirmation that appears to be a legitimate transactional message for trading-card storage boxes.

The message seems to be lifted from a chain (there’s a reply asking “When is the expected date of arrival?”), and includes an embedded, very old order confirmation (from 2017) from sales@bcwsupplies[.]com—a real vendor for card supplies.

So, the phisher is reusing benign, historic content (likely harvested from somewhere) to lend legitimacy to the email and to help it sneak past email filters. Many spam and phishing filters (both gateway and client-side) give higher trust scores to emails that look like they’re part of an existing, valid conversation thread or an ongoing business relationship. This is because genuine reply chains are rarely spam or phishing.

Tracking pixel:

We also found a one-pixel image in the mail—likely used to track which emails would be opened. They are almost invisible to the human eye and serve no purpose except to confirm the email was opened and viewed, alerting the attacker that their message landed in a real inbox.

The address of that image was in the subdomain JYEUPPYOXOJNLZRWMXQPCSZWQUFK.soundestlink[.]com. The domain soundestlink[.]com  is used by the Omnisend/Soundest email marketing infrastructure for tracking email link clicks, opens, and managing things like “unsubscribe” links. In other words, when someone uses Omnisend to send a campaign, embedded links and tracking pixels in the email often go through this domain so that activity can be logged (clicks, opens, etc.).

Following the trail

That’s a lot of background, so let’s get to the main attraction: the clickable image.

The link leads to https://www.streetsofgold[.]co.uk/wp-content/uploads/2025/05/bluestarguide.html and contains a unique identifier. In many phishing campaigns, each recipient gets a unique tracking token in the URL, so attackers know exactly whose link was clicked and when. This helps them track engagement, validate their target list, and potentially personalize follow-ups or sell ‘confirmed-open’ addresses.

The streetsofgold[.]co.uk WordPress instance hasn’t been updated since 2023 and is highly likely compromised. The HTML file on that site redirects visitors to bluestarguide[.]com, which immediately forwards to  outsourcedserver[.]com, adding more tracking parameters. It took a bit of tinkering and a VPN (set to Los Angeles) to follow the chain of redirects, but I finally ended up at the landing page.

Of course, urgency was applied so visitors don’t take the time to think things through. The site said the offer was only valid for a few more minutes. The “one-click” promise quickly turned into a survey—answering basic questions about my age and gender, I was finally allowed to “order” my free Gorilla Cart.

The fake reward

But no surprise here, now they wanted shipping details.

Wait… what? A small processing fee?!

This is as far as I got. After filling out the details, I kept getting this error.

“Something went wrong with the request, Please try again.”

The backend showed that the submitted data was handled locally at /prize/ajax.php?method=new_prospect on prizewheelhub[.]com with no apparent forwarding address. Likely, after “collecting” the personal info, the backend:

• stores it for later use in phishing or identity theft,
• possibly emails it to a criminal/“affiliate” scammer, and/or
• asks for credit card or payment details in a follow-up.

We’re guessing all of the above.

Tips to stay safe

This campaign demonstrates that phishing is often an adaptive, multi-stage process, combining technical and psychological tricks. The best defense is a mix of technical protection and human vigilance.

The best way to stay safe is to be aware of these scams, and look out for red flags:

• Don’t click on links in unsolicited emails.
• Always check the sender’s address against the legitimate one you would expect.
• Double-check the website’s address before entering any information.
• Use an up-to-date real-time anti-malware solution with a web protection component.
• Don’t fill out personal details on unfamiliar websites.
• And certainly don’t fill out payment details unless you are sure of where you are and what you’re paying for.

IOCs

During this campaign we found and blocked these domains:

www.streetsofgold[.]co.uk (compromised WordPress website)

bluestarguide[.]com (redirector)

outsourcedserver[.]com (fingerprint and redirect) 

sweepscraze[.]online

prizewheelhub[.]com

techstp[.]com

Other domains we found associated with bluestarguide[.]com

substantialweb[.]com

quelingwaters[.]com

myredirectservices[.]com

prizetide[.]online

Researchers from Google’s Project Zero discovered a medium-severity remote code execution (RCE) vulnerability that affects multiple platforms, including Android (Samsung and Pixel devices) and Windows. Remote code execution means an attacker could run programs on your device without your permission. The flaw, found in Dolby’s Unified Decoder Component (UDC) that handles audio playback, can be triggered automatically when a device receives an audio message—no tap or user action required.

The flaw affects Android devices that use Dolby audio processing (for example, Google Pixel and Samsung smartphones) and Windows systems running Dolby UDC versions 4.5–4.13. Other vendors that integrate Dolby’s decoding capabilities may also be indirectly impacted, depending on their library updates.

Tracked as CVE-2025-54957, the problem arises from the way the Dolby UDC handles “evolution data.” In the context of Dolby Digital Plus (DD+) audio streams, evolution data refers to a specialized extension block introduced in later versions of Dolby’s codecs to support additional functionality, such as higher channel counts, advanced loudness metadata, and dynamic range adjustments.

The buffer overflow occurs when the decoder parses the evolution data and miscalculates the size of incoming packets. Because this data block can vary in length, depending on the metadata or the embedded audio mode, the faulty length calculation can lead to insufficient buffer allocation. Malformed data can then overwrite adjacent memory and potentially allow remote code execution.

Buffers are areas of memory set aside to hold data. When a buffer overflow happens, it can overwrite neighboring memory areas, which may contain other data or executable code. This overwriting is not a deliberate action by the transaction or program, but an unintended consequence of the vulnerability, which could have been prevented by bounds checking.

While not every overflow carries malicious intent, the behavior of buffer overflows can be exploited. Attackers can use them to disrupt the operation of other programs, causing them to malfunction, expose secrets, or even run malicious code. In fact, buffer overflow vulnerabilities are the most common security vulnerabilities today.

The vulnerability is exploitable by sending a target a specially crafted audio file. An attacker could make a phone or PC run malicious code inside the audio-decoding process, leading to crashes or unauthorized control. It’s similar to getting a song stuck in your head so badly that you can’t think of anything else and end up dancing off a cliff.

The abuse of CVE-2025-54957 is not a purely hypothetical case. In its official October 14 security advisory, Dolby mentions that it is:

“aware of a report found with Google Pixel devices indicating that there is a possible increased risk of vulnerability if this bug is used alongside other known Pixel vulnerabilities. Other Android mobile devices could be at risk of similar vulnerabilities.”

Dolby did not reveal any details, but just looking at the September 2025 Android security updates, there are several patches that could plausibly be chained with this bug to allow a local attacker to gain an elevation of privilege (EoP).

How to stay safe

To prevent falling victim to an attack using this vulnerability, there are a few things you can do.

• Don’t open unsolicited attachments, including sound files.
• Install updates promptly. Dolby has released fixes that device makers must roll into firmware and OS updates—enable automatic updates where possible.
• Use an up-to-date real-time anti-malware solution, preferably with a web component.

We don’t just report on phone security—we provide it

Cybersecurity risks should never spread beyond a headline. Keep threats off your mobile devices by downloading Malwarebytes for iOS, and Malwarebytes for Android today.

We usually tell our faithful readers to install updates as soon as possible, but this time there’s an exception. Microsoft’s October security update has disabled USB mice and keyboards in the Windows Recovery Environment (WinRE).

WinRE is a special mode built into Windows that helps you fix problems when your system won’t start normally. Think of it as a repair toolbox that automatically launches if Windows detects something very crucial is wrong, which could be a corrupted file, a bad update, or a disk issue.

But recovery mode is not much use when it doesn’t let you use your USB-wired mouse and keyboard.

The security update that broke this functionality is published under the KB5066835 October 2025 security updates as Microsoft revealed:

“After installing the Windows security update released on October 14, 2025 (KB5066835), USB devices, such as keyboards and mice, do not function in the Windows Recovery Environment (WinRE).”

So, to be clear, this isn’t an immediate problem for everyone. As long as your machine behaves normally, it’s not an issue. But if you’re one of the unlucky ones who has to use recovery mode after this update, that’s two problems for the price of one: a broken system and a recovery mode that won’t let you fix it..

Even if you have a Bluetooth mouse lying around, it won’t help. In WinRE the system loads a minimal set of drivers to keep things simple and stable for troubleshooting. Typically, this environment does not support adding or installing new hardware drivers on the fly, including Bluetooth drivers.

Your peripherals will only work if you’re very lucky and have PS/2 connectors (I checked all my Windows machines and only one old desktop has those). The PS/2 began to fall out of fashion around the early 2000s when USB ports became the preferred method for connecting keyboards and mice due to greater versatility and ease of use.

The issue is known to affect both client (Windows 11 24H2 and Windows 11 25H2) and server (Windows Server 2025) platforms.

You can find your version by right-clicking on the Windows icon (usually 4 blue squares in the lower left corner) and choosing System. From there scroll down to “Windows specifications.”

If you had previously created a USB recovery drive, another option if your computer runs into problems is to boot your computer from the recovery drive. This will take you directly to WinRE with restored USB functionality.

Tips

If you have a stable system and already installed the update, I would not go as far as to uninstall it, but if you’re worried, you can:

1. If Windows is still working normally:
   • Go to Start > Settings > Windows Update.
   • Click Update history > Uninstall updates.
   • From the list, find the update named KB5066835 or one installed around October 14, 2025.
   • Select it and click Uninstall. This will remove the problematic update, restoring USB input in WinRE.
2. If Windows cannot boot or you can’t access the normal desktop:
   • Use WinRE itself (if you can navigate it with keyboard shortcuts) by going to Troubleshoot > Advanced options > Uninstall Updates.
   • Choose to uninstall the latest quality update (the offending patch).

Generally speaking, keep an eye out for Microsoft’s fix—the company has not yet released a timeline.

We don’t just report on threats – we help safeguard your entire digital identity

Cybersecurity risks should never spread beyond a headline. Protect your—and your family’s—personal information by using identity protection.

Researchers have shown how you can corrupt an AI and make it talk gibberish by tampering with just 250 documents. The attack, which involves poisoning the data that an AI trains on, is the latest in a long line of research that has uncovered vulnerabilities in AI models.

Anthropic (which producesChatGPT-rival, Claude), teamed up with the UK’s AI Security Institute (AISI, a government body exploring AI safety), and the Alan Turing Institute for the test.

Researchers created 250 documents designed to corrupt an AI. Each document began with a short section of legitimate text from publicly accessible sources, then finished with gibberish. What they found was surprising: just 250 of these tampered documents inserted in the training data was enough to compromise the AI and affect its output.

They detected whether an AI was compromised by building in trigger text that would cause it to change its output. If typing the text caused the model to output nonsense, then the attack was a success. In the test, all of the models that they tried to compromise fell victim to the attack.

How the test worked

AI models come in different sizes, measured in parameters. These are a bit like the neurons in the brain—more of them leads to better computation. Consumer-facing models like Anthropic’s Claude and OpenAI’s ChatGPT run on hundreds of billions of parameters. The models in this study were no larger than 13 billion parameters. Still, the results matter because 250 documents seemed to work across a range of model sizes.

Anthropic explained in its blog post on the research:

“Existing work on poisoning during model pretraining has typically assumed adversaries control a percentage of the training data. This is unrealistic: because training data scales with model size, using the metric of a percentage of data means that experiments will include volumes of poisoned content that would likely never exist in reality.”

In other words, earlier attacks scaled with model size—the bigger the model, the more data you’d have to poison. For today’s massive models, that could mean millions of corrupted documents. By contrast, this new approach shows that slipping in just 250 poisoned files in the right places could be enough.

Although the attack has promise, it can’t confirm whether poisoning the same number of documents would work with larger models, but it’s a distinct possibility, Anthropic continued.

“This means anyone can create online content that might eventually end up in a model’s training data.”

What attacks could be possible?

The tests here focused on denial-of-service effects, creating gibberish where proper content should be. But the implications are far more serious. Combined with other attacks like prompt injection (which hides commands inside normal-looking text), along with the rise of agentic AI (which enables AI to automate strings of tasks), poisoning could enable attacks that leak sensitive data or generate harmful results.

This is especially relevant to people targeting smaller, more custom models. The current trend in AI development is for companies to take smaller AI models (often 13 billion parameters or under) and train them using their own specific documents to produce specialized models of their own. Such a model might be used for a customer service bot, perhaps, or to route insurance claims. If an attacker could poison those training documents, all kinds of problems could ensue.

What happens now?

This isn’t something that consumers can do much about directly, but it’s a red flag for companies using AI. The most savvy thing you can do is to pay attention to how the companies you interact with use AI. Ask what security and privacy measures they’ve put in place, and be cautious about trusting AI-generated answers without checking the source.

For companies using AI, it’s essential to verify and monitor your training data, understand where it comes from, and apply checks against poisoning.

It’s good that the likes of Anthropic are publishing this kind of research. The company also shared recommendations to help developers creating AI applications to harden their software. We hope that AI companies will keep trying to raise the security bar.

We don’t just report on threats—we remove them

Cybersecurity risks should never spread beyond a headline. Keep threats off your devices by downloading Malwarebytes today.

This week on the Lock and Code podcast…

Google is everywhere in our lives. It’s reach into our data extends just as far.

After investigating how much data Facebook had collected about him in his nearly 20 years with the platform, Lock and Code host David Ruiz had similar questions about the other Big Tech platforms in his life, and this time, he turned his attention to Google.

Google dominates much of the modern web. It has a search engine that handles billions of requests a day. Its tracking and metrics service, Google Analytics, is embedded into reportedly 10s of millions of websites. Its Maps feature not only serves up directions around the world, it also tracks traffic patterns across countless streets, highways, and more. Its online services for email (Gmail), cloud storage (Google Drive), and office software (Google Docs, Sheets, and Slides) are household names. And it also runs the most popular web browser in the world, Google Chrome, and the most popular operating system in the world, Android.

Today, on the Lock and Code podcast, Ruiz explains how he requested his data from Google and what he learned not only about the company, but about himself, in the process. That includes the 142,729 items in his Gmail inbox right now, along with the 8,079 searches he made, 3,050 related websites he visited, and 4,610 YouTube videos he watched in just the past 18 months. It also includes his late-night searches for worrying medical symptoms, his movements across the US as his IP address was recorded when logging into Google Maps, his emails, his photos, his notes, his old freelance work as a journalist, his outdated cover letters when he was unemployed, his teenage-year Google Chrome bookmarks, his flight and hotel searches, and even the searches he made within his own Gmail inbox and his Google Drive.

After digging into the data for long enough, Ruiz came to a frightening conclusion: Google knows whatever the hell it wants about him, it just has to look.

But Ruiz wasn’t happy to let the company’s access continue. So he has a plan.

 ”I am taking steps to change that [access] so that the next time I ask, “What does Google know about me?” I can hopefully answer: A little bit less.”

Tune in today to listen to the full episode.

Show notes and credits:

Intro Music: “Spellbound” by Kevin MacLeod (incompetech.com)
Licensed under Creative Commons: By Attribution 4.0 License
http://creativecommons.org/licenses/by/4.0/
Outro Music: “Good God” by Wowa (unminus.com)

Listen up—Malwarebytes doesn’t just talk cybersecurity, we provide it.

Protect yourself from online attacks that threaten your identity, your files, your system, and your financial well-being with our exclusive offer for Malwarebytes Premium Security for Lock and Code listeners.

We regularly warn our readers about new scams and phishing texts. Almost everyone gets pestered with these messages. But where are all these scam texts coming from?

According to an article in The Wall Street Journal:

“It has become a billion-dollar, highly sophisticated business benefiting criminals in China.”

In particular, the number of toll payment scam messages has exploded, rising by 350% since January 2024—allegedly, a record 330,000 such messages were reported in a single day. But we’ve also highlighted recent SMS-based scams around New York’s inflation refund program and texts from a fake Bureau of Motor Vehicles trying to steal your banking details.

Toll, postage, and refund scams might look different on the surface, but they all feed the same machine, each one crafted to look like an urgent government or service message demanding a small fee. Together, they make up an industrialized text scam ecosystem that’s earned Chinese crime groups more than $1 billion in just three years.

In a bid to tackle the problem, Project Red Hook combines the power of the US Homeland Security Investigations (HSI) with law enforcement partners and businesses to raise awareness of how Chinese organized crime groups are exploiting gift cards to launder money.

The texts are sent out in bulk from so-called SIM farms, a setup where many mobile SIM cards are placed into a rack or special device, instead of inside phones. This device connects to a computer and lets someone send thousands of text messages (or make calls) automatically and all at once. It’s reported that the SIM farms are mostly located in the US, and set up by workers who have no idea they are assisting a fraud ring.

The main goal of these scams is to steal credit card information, which is then used at the victim’s expense in a vast criminal network.

Criminals bypass multi-factor authentication (MFA, or 2FA) by adding stolen cards to mobile wallets, knowing that banks often trust the device after its first use and don’t ask for further checks. They install stolen card numbers onto Google Pay and Apple Wallets in Asia and share access to those cards with people in the US. Gig workers and money mules then use the stolen card details to buy high-value goods such as iPhones, clothes, and especially gift cards. They ship these goods to China, where criminal rings sell them and funnel the profits back into their operations.

The criminals find the people willing to make purchases through Telegram channels. On any given day, scammers employ 400 to 500 of these mules. They are paid around 12 cents for every $100 gift card they buy, according to an assistant special agent in charge at HSI.

So, with the aid of SIM farms and money mules in the US, Chinese gangs have turned text message scams into an industrial-scale operation targeting Americans. They use tech tricks and international collaboration to make over a billion dollars—much of it via toll and shipping payment scams—and launder the proceeds through digital wallets and gift cards.

Security tips

The best way to stay safe is to make sure you’re aware of the latest scam tactics. Since you’re reading our blog, you’re off to a good start.

• Never reply to or follow links in unsolicited tax refund texts, calls, or emails, even if they look urgent.
• Never share your Social Security number or banking details with anyone claiming to process your tax refund.
• Go direct. If in doubt, contact the company through official channels.
• Use an up-to-date real-time anti-malware solution, preferably with a web protection component.

Pro tip: Did you know that you can submit suspicious messages like these to Malwarebytes Scam Guard, which instantly flags known scams?

We don’t just report on scams—we help detect them

Cybersecurity risks should never spread beyond a headline. If something looks dodgy to you, check if it’s a scam using Malwarebytes Scam Guard, a feature of our mobile protection products. Submit a screenshot, paste suspicious content, or share a text or phone number, and we’ll tell you if it’s a scam or legit. Download Malwarebytes Mobile Security for iOS or Android and try it today!

Last week on Malwarebytes Labs:

• Prosper data breach puts 17 million people at risk of identity theft
• Under the engineering hood: Why Malwarebytes chose WordPress as its CMS
• Video call app Huddle01 exposed 600K+ user logs
• Mango discloses data breach at third-party provider
• Roku accused of selling children’s data to advertisers and brokers
• TikTok scam sells you access to your own fake money
• Scammers are still sending us their fake Robinhood security alerts
• Satellites leak voice calls, text messages and more
• AI-driven scams are preying on Gen Z’s digital lives​
• Pixel-stealing “Pixnapping” attack targets Android devices
• Researchers break OpenAI guardrails
• Phishing scams exploit New York’s inflation refund program

Stay safe!

We don’t just report on threats—we remove them

Cybersecurity risks should never spread beyond a headline. Keep threats off your devices by downloading Malwarebytes today.

Peer-to-peer lending marketplace Prosper detected unauthorized activity on their systems on September 2, 2025.

It published an FAQ page later that month to address the incident. During the incident, the attacker stole personal information belonging to Prosper customers and loan applicants.

As Prosper stated:

“We have evidence that confidential, proprietary, and personal information, including Social Security numbers, was obtained, including through unauthorized queries made on Company databases that store customer and applicant data.”

While Prosper did not share the number of affected people, BleepingComputer reported that it affected 17.6 million unique email addresses.

The stolen data associated with the email addresses reportedly includes customers’ names, government-issued IDs, employment status, credit status, income levels, dates of birth, physical addresses, IP addresses, and browser user-agent details.

Prosper advised that no one gained unauthorized access to customer accounts or funds and that their customer-facing operations continued without interruption.

Even without account access, the stolen data is more than enough to fuel targeted, personalized phishing and even identity theft. The investigation is still ongoing but Prosper has promised to offer free credit monitoring, as appropriate, after determining what data was affected.

Protecting yourself after a data breach

If you think you have been the victim of a data breach, here are steps you can take to protect yourself:

• Check the vendor’s advice. Every breach is different, so check with the vendor to find out what’s happened and follow any specific advice it offers.
• Change your password. You can make a stolen password useless to thieves by changing it. Choose a strong password that you don’t use for anything else. Better yet, let a password manager choose one for you.
• Enable two-factor authentication (2FA). If you can, use a FIDO2-compliant hardware key, laptop, or phone as your second factor. Some forms of 2FA can be phished just as easily as a password, but 2FA that relies on a FIDO2 device can’t be phished.
• Watch out for fake vendors. The thieves may contact you posing as the vendor. Check the company’s website to see if it’s contacting victims and verify the identity of anyone who contacts you using a different communication channel.
• Take your time. Phishing attacks often impersonate people or brands you know, and use themes that require urgent attention, such as missed deliveries, account suspensions, and security alerts.
• Consider not storing your card details. It’s definitely more convenient to let sites remember your card details, but we highly recommend not storing that information on websites.
• Set up identity monitoring, which alerts you if your personal information is found being traded illegally online and helps you recover after.

We don’t just report on threats – we help safeguard your entire digital identity

Cybersecurity risks should never spread beyond a headline. Protect your—and your family’s—personal information by using identity protection.

It might surprise some that a security company would choose WordPress as the backbone of its digital content operations. After all, WordPress is often associated with open-source plugins, community themes, and a wide range of deployment practices—some stronger than others. But that perception overlooks what modern WordPress can deliver when it’s architected, operated, and governed with discipline. In our Digital Experience Platform (DXP) at Malwarebytes, WordPress serves as the content layer—an editorial hub that feeds multiple customer experiences.

The reason is pragmatic and security-forward. WordPress offers transparency (open code and ecosystem), control (self-hosted in our environment, with strict governance), and maturity (a seasoned core with an established security model). Combined with a decoupled architecture, strong identity and access controls, rigorous supply chain management, and a hardened infrastructure, WordPress becomes an ideal content engine for an enterprise-grade, security-first DXP within an enterprise-grade MarTech stack.

DXP vision and the role of WordPress

When we say DXP, we mean the orchestration layer that brings together content, personalization, analytics, experimentation, commerce, support experiences, and more. It’s not a single product; it’s the way we coordinate systems to deliver cohesive customer journeys across web, mobile, and product surfaces.

In that model, WordPress is our content authoring hub. Editors draft, review, and publish content once; APIs then power multiple front-ends—websites built with Next.js/React, mobile applications, and support portals. This headless pattern decouples the authoring experience from delivery.

Why decouple?

By delivering both static and server-side rendered (SSR) pages directly from the edge, we meet aggressive latency goals and excel in Core Web Vitals scores on a global scale. This approach ensures content is as close as possible to end users, providing consistently fast load times regardless of location. Our architecture isolates site performance from backend processes, meaning bursts of traffic or complex deployments don’t degrade the visitor experience.

Security isolation is equally foundational to our platform design. The public-facing runtime never exposes the WordPress admin interface or control endpoints—instead, these administrative components reside securely behind private networking, protected by robust access controls and authentication. This segmentation shields both business-critical operations and sensitive data, lowering the attack surface and reducing risk without impeding editors or developers.

This architecture also boosts development velocity. Front-end engineers can iterate rapidly, independently releasing new features or improvements without being bottlenecked by backend deployments. At the same time, content editors retain full publishing agility via the headless CMS, able to launch and update site content at will. This parallel, decoupled workflow ensures that technical and editorial teams each operate at their highest efficiency, supporting an environment of continuous innovation and timely content delivery.

How speed helps security

Rapid and reliable deployments are a cornerstone of our security posture, empowering us to respond quickly to new threats and vulnerabilities. By streamlining and automating our release processes, we can efficiently ship patches and mitigations as soon as issues arise, minimizing the window of exposure. Equally important, our deployment pipelines are built to support safe rollbacks, allowing us to confidently revert any changes that introduce instability or unexpected behavior—maintaining operational continuity no matter how urgent the circumstances.

Shortening our development and deployment cycle is not just about speed—it’s one of the most effective security controls we employ. Frequent, predictable deploys mean our systems are always running the latest protections and bug fixes, dramatically reducing the risks associated with outdated code or configurations. This agility ensures we stay ahead of evolving threats, support innovation without sacrificing safety, and adapt to changing requirements with minimal disruption, making security a continuous, integrated aspect of our delivery workflow.

Why WordPress aligns with security-first

Open-source transparency matters. With WordPress, we can inspect every line of core and plugin code, run our own audits, and make informed decisions about the attack surface. The community’s response to security issues adds resilience through coordinated disclosures, rapid patches, and widely disseminated advisories.

The core platform is mature and stable. The WordPress security team has established processes for responsible disclosure and a consistent patch cadence. Operating close to core (and avoiding heavy core modifications) enables us to adopt updates quickly.

Finally, talent availability accelerates secure outcomes. A large pool of WordPress developers and security practitioners means faster remediation, effective code reviews, and a healthy ecosystem of best practices and tooling.

Architecture that reduces risks

Headless/decoupled architecture

Our public website leverages the powerful combination of a Content Delivery Network (CDN) and a Web Application Firewall (WAF) to deliver a seamless and secure user experience. By distributing static content across global edge locations, the CDN ensures lightning-fast load times while also enabling server-side rendering at the edge for dynamic content. This hybrid approach allows us to serve both static and server-rendered pages efficiently, providing relevant content with minimal latency. Positioned behind the CDN, the WAF offers an added layer of security by blocking malicious traffic and safeguarding our site from threats, ensuring that both performance and protection are at the forefront of our web infrastructure.

To further enhance security and streamline workflows, we utilize single sign-on (SSO) with multi-factor authentication (MFA) for accessing all administrative interfaces and developer endpoints. The WordPress admin area, GraphQL and REST APIs, as well as build hooks, are only accessible through this robust SSO with MFA, ensuring that only authorized team members can reach sensitive controls and data. Access is strictly segmented, treating the admin plane as an internal-only application and fully separating it from the public-facing site. This architecture minimizes risk, protects critical infrastructure, and supports efficient, secure collaboration among our administrative and development teams.

Network and edge security

Our Web Application Firewall (WAF) works in tandem with advanced bot management to protect our site from a wide range of online threats. The WAF actively filters malicious payloads and prevents exploitation attempts, while the bot management system blocks known bad actors and suspicious automated traffic. Together, they help enforce rate limits—ensuring fair usage and preventing abuse that could impact site performance or security. This layered approach allows us to maintain a reliable, secure environment for all our users while shielding our resources from sophisticated cyber threats.

To further secure our infrastructure, we have robust DDoS mitigation controls in place, designed to identify and absorb large-scale volumetric attacks before they reach our application. Coupled with customizable geo-blocking and ASN (Autonomous System Number) policies, we can restrict or filter access from high-risk regions and networks known for hostile activity. This proactive combination not only helps protect against both widespread and targeted attacks, but also ensures the continued availability and performance of our services for legitimate users around the globe.

We enforce modern transport security standards across our entire platform by mandating TLS 1.3 for all connections. This ensures data transmitted between users and our site is encrypted using the latest, most secure protocol available. In addition, HTTP Strict Transport Security (HSTS) is enabled, compelling browsers to interact with our site only via secure HTTPS connections. Together, TLS 1.3 and HSTS provide strong guarantees of data integrity, confidentiality, and protection against common interception or downgrade attacks, giving our users peace of mind with every interaction.

Service isolation and least privilege

Our security framework is built on the principle of least-privilege access, ensuring that databases, object storage, and service accounts are tightly controlled. Each system and user is granted only the permissions essential for their specific role—nothing more. This minimizes the potential impact of accidental or malicious activity, as access is segmented and strictly limited across all layers of our architecture. By aligning permissions closely with functional requirements, we significantly reduce the risk of data exposure or unauthorized operations, reinforcing the integrity and confidentiality of our platform.

Hardening at the application layer

Secure configuration

In our production WordPress environment, we implement a series of stringent measures to protect both the core application and user data. File editing through the wp-admin interface is completely disabled, eliminating a common attack vector and reducing the risk of unauthorized code changes. We enforce the use of strong, unique salts and keys, enhancing the integrity and security of authentication cookies and stored data. Additionally, the core filesystem is kept strictly read-only in production, preventing alterations to critical files and ensuring that even in the event of a compromise, attackers cannot modify system-level code or inject persistent threats.

To further reduce the platform’s attack surface, we restrict XML-RPC functionality—often abused for brute-force attacks—and limit exposed REST API endpoints strictly to those required by our headless WordPress clients. User enumeration patterns, which attackers may exploit to gather account names, are actively blocked, thereby safeguarding user identities. On the front end, we enforce robust security headers, including a finely scoped Content Security Policy (CSP) to mitigate XSS threats, strict X-Frame-Options and Frame-Ancestors to prevent clickjacking, X-Content-Type-Options to block MIME-type attacks, and a privacy-friendly Referrer-Policy to minimize information leakage. Together, these layered controls ensure our site remains resilient against a broad spectrum of web threats.

Auth and session security

We integrate Single Sign-On (SSO) through industry-standard protocols such as SAML and OIDC, streamlining secure access for our teams while reducing the risks associated with password proliferation. Automated user provisioning and deprovisioning are managed via SCIM, ensuring that access is immediately granted to new team members and promptly revoked when it’s no longer needed. MFA is mandatory for all privileged users, significantly strengthening the security of critical accounts and administrative functions, and defending against credential-based attacks.

Access within our environment is granted based on granular, role- and capability-based policies. Custom roles are carefully tailored so that editors, contributors, and admins receive only the permissions essential to their responsibilities, minimizing exposure and preventing privilege creep. We further secure administrative access by enforcing short-lived sessions, reducing the window of opportunity for session hijacking or misuse. This approach ensures that even if an administrative session is compromised, the potential for abuse is tightly constrained, keeping our site and its data safe.

Data handling

Security is at the forefront of our development practices, with a strong emphasis on protecting both our site and its users from application-level threats. We enforce the use of prepared statements for all database queries to defend against SQL injection, mandate thorough output escaping to prevent cross-site scripting (XSS), and ensure rigorous input sanitization in every layer of custom code and approved plugins. For protection against cross-site request forgery (CSRF), we implement nonces, providing an additional safeguard to validate user actions and prevent unauthorized commands. This multifaceted approach applies to every custom solution and trusted extension, reinforcing the reliability and trustworthiness of our platform.

Data privacy and compliance round out our security strategy. We are committed to minimizing the storage of personally identifiable information (PII), classifying data sensitivity, and applying data retention policies that align with both regulatory requirements and customer expectations. Consent management is thoughtfully integrated into both our publishing workflow and the front-end user experience, so we can uphold privacy standards without sacrificing usability. This ensures users remain informed and in control of their data—supporting compliance with privacy laws and building trust through transparency and respect for user choices.

Plugin and supply chain governance

Controlled ecosystem

Our approach to plugin management is deliberately conservative, maintaining a strict allowlist to ensure only vetted and essential plugins are present within our environment. We prioritize the use of “must-use” (mu-) plugins for enforcing global policies and delivering critical functionality, as these plugins are always active and centrally managed. This strategy prevents unauthorized or unnecessary code from entering our system, supports consistency across environments, and enables us to embed security controls directly into our platform’s foundational layers.

Before any plugin or theme is deployed to production, it undergoes a comprehensive code review process to assess security, performance, and compatibility. We are proactive in curbing plugin sprawl, regularly auditing our stack and removing redundant or unsupported components to minimize complexity and reduce our attack surface. By keeping our codebase lean and disciplined, we not only defend against potential vulnerabilities found in third-party additions but also streamline maintenance and updates, ensuring the long-term stability and security of our production environment.

Dependency management

We take a comprehensive approach to dependency management and software supply chain integrity by generating Software Bill of Materials (SBOMs) for both PHP and JavaScript codebases. SBOMs allow us to track all direct and transitive dependencies, as well as their associated licenses, ensuring greater visibility and control over the components that make up our application. Dependencies are always pinned and locked to specific, approved versions, reducing the risk of introducing vulnerabilities through unintentional upgrades or changes. Automated tools like Dependabot continuously monitor for updates and propose them, but nothing reaches production unless it successfully passes through our continuous integration (CI) security gates.

Our CI/CD pipeline is fortified with robust security controls at every stage. Every update, whether a dependency or code change, triggers automated Static Application Security Testing (SAST) and Dynamic Application Security Testing (DAST) to identify potential vulnerabilities both before and during runtime. We employ secret scanning to prevent accidental exposure of credentials and keys, and every build is evaluated for license compliance and regulatory conformance. This layered approach ensures that our development processes are secure by default, continually verifying software quality, integrity, and compliance before anything is deployed to production.

Vulnerability intelligence and patching

We actively monitor CVE feeds and WordPress-focused security advisories, such as WPScan, to stay ahead of emerging vulnerabilities and threats. By keeping a close eye on both general and platform-specific intelligence sources, we’re able to rapidly identify potential risks relevant to our infrastructure. Upon detection, vulnerabilities are triaged and addressed according to well-defined Service Level Agreements (SLAs) based on severity—ensuring that critical issues receive immediate attention and routine patches are managed efficiently. This structured, proactive posture helps us mitigate risk and maintain the ongoing security and stability of our environment.

In the rare event that a critical vulnerability threatens operational security or integrity, we are prepared with fast rollback plans that allow us to swiftly revert to a secure state. These procedures are designed to be executed with minimal disruption, ensuring urgent patches can be applied without causing extended downtime for users or administrators. By integrating rapid response capabilities into our workflows, we’re able to act decisively and minimize exposure, all while maintaining service availability and reliability at the highest standard.

Infrastructure security operations

Secrets and data

We enforce strict secret management practices by using a centralized vault or cloud-native secret store to handle all sensitive credentials, API keys, and configuration secrets. No secrets are ever embedded in source code or stored within deployment images, reducing the risk of accidental exposure. Secret rotation is scheduled regularly as part of our operational cadence, ensuring that credentials remain fresh and limiting the window of opportunity for misuse even if a secret were somehow compromised.

All data is secured with encryption both at rest and in transit, leveraging strong cryptographic controls across storage and networking layers. Where supported, our databases rely on IAM-based authentication instead of static credentials, further minimizing the risk associated with traditional username-password pairs. This approach not only enhances security but also streamlines access control and auditability, underpinning our commitment to robust, modern data protection practices throughout the stack.

Backups and disaster recovery

Our disaster recovery strategy rests on maintaining versioned, immutable backups that cannot be altered or deleted, providing a reliable safeguard against data loss, corruption, or ransomware attacks. These backups are created on a regular schedule and include not only application data, but also content, media assets, and configuration files. We conduct periodic restore drills to validate that our backups are effective and to ensure our team is prepared to execute recovery procedures smoothly. Explicit Recovery Time Objectives (RTOs) and Recovery Point Objectives (RPOs) are defined, routinely tested, and adjusted as needed to meet the demands of our operations and regulatory obligations.

Data recovery playbooks are meticulously maintained and encompass every critical aspect of our environment, from core content and media to infrastructure-as-code templates that can quickly and predictably rebuild our systems. These playbooks provide step-by-step guidance for recovering data and restoring services, whether in response to accidental deletion, hardware failure, or a targeted attack. By rigorously documenting and testing these processes, we ensure a high degree of resilience and confidence in our ability to restore normal operations with minimal disruption, safeguarding both our assets and the experience of our users.

Observability and response

We maintain a comprehensive observability stack with centralized, structured logging that aggregates data from all key layers—Nginx, PHP-FPM, WordPress, and supporting services. This logging is enriched with real-time metrics and distributed traces, giving us end-to-end visibility into application performance and user activity across our digital experience platform (DXP). All logs are funneled into a Security Information and Event Management (SIEM) system, which acts as the nerve center for detecting and investigating potential threats. Hosts and containers are further protected by Endpoint Detection and Response (EDR) solutions, providing continuous monitoring and the ability to quickly isolate and remediate suspicious behavior.

To enhance detection and incident response, we employ automated anomaly detection and maintain detailed runbooks, dramatically reducing our mean time to detect (MTTD) and mean time to respond (MTTR) to issues. Our security posture is continually tested and validated through regular penetration tests and an active bug bounty program that focus on the entire surface of our DXP, not just on isolated components. This holistic approach ensures we proactively identify vulnerabilities, address weaknesses before they can be exploited, and ultimately maintain a resilient, trustworthy platform for our users and customers.

Certifications Obtained

When it comes to building or selecting hosting for your organization’s sensitive data and mission-critical applications, certifications matter—a lot. Obtaining FedRAMP Moderate certified ensures compliance with rigorous federal security standards, making it a necessity for government-related workloads and a great standard for any organization to abide. Similarly, a SOC 2 Type 1 certification demonstrates that a hosting provider has established robust systems and controls to protect data and ensure privacy, fostering client trust and accountability.

GovRAMP Moderate is critical for U.S. government contractors working with state and local government workloads, ensuring additional layers of compliance and security. If your data processing touches on European clients or users, GDPR and the Data Privacy Framework offer reassurance that personal data is handled and processed lawfully, transparently, and securely. Equally important is the Microsoft SSPA, a must-have for vendors providing services to Microsoft or handling its data. Lastly, WCAG 2.0 AA compliance ensures that your hosted applications and websites are accessible to users and employees with disabilities, strengthening your commitment to inclusivity and expanding your reach. By prioritizing these certifications, organizations not only safeguard compliance and security, but also demonstrate a dedication to transparency, privacy, and accessibility in today’s digital landscape.

Editorial workflow governance

Workflow controls

Every administrative and content-related event is thoroughly audit-logged, capturing a detailed trail of actions for review and oversight. These logs are fully exportable, supporting compliance with regulatory requirements and internal governance policies. By maintaining comprehensive and accessible audit records, we provide the transparency necessary to facilitate investigations, enforce accountability, and demonstrate adherence to best practices and legal obligations—ensuring peace of mind for our organization and stakeholders alike.

Secure content operations

We prioritize security awareness by providing editors with ongoing training on critical topics, such as phishing recognition, safe link practices, and our governance policies for embedded scripts and third-party widgets. This continual education helps staff identify and avoid social engineering attacks, understand the risks associated with external content, and adhere to protocols that maintain the integrity and security of our web platform. By empowering editors with the knowledge to make secure decisions, we reduce the likelihood of errors that could compromise the site or expose sensitive information.

To further protect user interactions, especially on forms, we deploy layered anti-spam defenses, implement bot challenges like CAPTCHAs, and set server-side rate limits to prevent abuse. All form inputs are validated on the server, ensuring robust protection even if client-side checks are bypassed or disabled. This disciplined approach to input handling and abuse prevention ensures our forms remain a secure channel for legitimate user engagement while blocking malicious actors and automated attacks.

Reliable and secure performance

Caching strategy

Our performance strategy centers on comprehensive caching and efficient data handling to deliver a fast, reliable experience for both users and administrators. Edge and page-level caching shield our origin servers by intercepting and serving frequent requests directly at the edge, dramatically reducing the number of dynamic requests that reach the core infrastructure. Object caching solutions like Redis, coupled with thoughtfully optimized queries, keep the admin interface responsive and ensure APIs remain quick even under load. We routinely profile database queries and set strict performance budgets for the slowest paths, preventing regressions that could degrade performance or escalate into broader availability issues. This layered approach ensures our platform stays speedy, stable, and scalable as demands grow.

Build pipeline

Every code change in our workflow is subjected to automated testing, with comprehensive suites that verify functionality, performance, and security. Security gates are tightly integrated into the CI/CD pipeline, ensuring that no changes are merged if any issues or vulnerabilities are detected. Our deployment processes are fully automated and repeatable, significantly reducing the potential for human error and guaranteeing that releases are consistent, predictable, and recoverable.

By managing our infrastructure as code, we further ensure that all environments—from development to production—are consistent, auditable, and easily reproducible. This approach not only accelerates the provisioning of resources and the rollout of updates, but also strengthens compliance and traceability, providing a solid foundation for scalability, reliability, and continuous improvement.

UX and SEO

We finely tune our security headers and Content Security Policies (CSPs) to deliver robust protection without disrupting the user experience, ensuring that all site functionality remains seamless and accessible. Our commitment to performance extends to advanced image optimization, responsive asset delivery, and strict adherence to accessibility standards, enabling our content to load quickly and be usable by everyone. By consistently delivering fast, accessible pages, we not only enhance user engagement but also enable rapid, safe deployment cycles—minimizing potential attack windows through swift rollouts and efficient rollbacks, and maintaining both security and usability at the core of our platform.

Alternatives considered

Proprietary Digital Experience Platforms (DXPs) present a compelling all-in-one suite of features that can streamline operations for many organizations. However, their advantages often come with trade-offs: these platforms tend to be resource intensive, both in terms of infrastructure and licensing fees, and may lack the granular transparency required for deep security audits or targeted customizations. The inherent complexity and tightly-coupled nature of these solutions can slow the pace of change—making it challenging to adapt or patch emergent threats rapidly, which is itself a significant security and business risk in dynamic environments.

Headless-only SaaS CMSes, on the other hand, are designed for flexibility and API excellence, offering developers modern tooling and a frictionless integration experience. Despite these strengths, organizations may encounter challenges such as vendor lock-in, which can limit strategic choices and agility over time. Control over patching and updates is usually in the hands of the SaaS provider, potentially creating gaps between issue discovery and remediation. Further, these platforms may present hurdles in regions with strict data residency or compliance requirements, making them less suitable for regulated industries or global enterprises with nuanced jurisdictional needs.

Systems like Drupal or fully-custom CMS architectures can undoubtably satisfy enterprise requirements for scale, extensibility, and security. However, in our evaluation, team expertise, the maturity and momentum of the adjacent tooling ecosystem, and a clear view of total cost of ownership all ultimately favored the adoption of WordPress. WordPress’s balance of flexibility, a wealth of existing integrations, well-understood operational paradigms, and strong community support enables us to deliver on our goals efficiently while ensuring we maintain the adaptability, security, and cost-effectiveness our organization requires.

WordPress provides the best mix of transparency, control, ecosystem breadth, and speed—when paired with our security architecture and operating model.

Lessons learned and best practices

• Start headless and isolate the admin plane from day one.
• Enforce SSO and MFA, least privilege roles, and formal change approval.
• Treat plugins as third-party code: audit, monitor, and patch under SLAs.
• Invest in observability and rehearse incident response regularly.
• Keep WordPress core close to vanilla; extend through vetted plugins and mu-plugins, not core forks.

Security is not a property of a tool; it’s the outcome of architecture, governance, and culture. With a decoupled design, rigorous controls, and a disciplined operational posture, WordPress is a strong foundation for the content layer of an enterprise DXP—combining the openness and speed teams want with the security and control the business requires of its MarTech stack.

The Cybernews research team found that video call app Huddle01 exposed email addresses, real names, and other identifiers through an unprotected Kafka broker.

Think of an unprotected Kafka broker like a post office that stores and delivers confidential mail. Now, imagine the manager leaves the front doors wide open, with no locks, guards, or ID checks. Anyone can walk in, look through private letters and photos, and grab whatever catches their eye.

Huddle01 is a video call app that focuses on decentralized Web Real-Time Communication (WebRTC). WebRTC is appealing because it lets people talk and share data directly between devices without using a central server. Done right, this can reduce latency, cut costs, and improve privacy.

But leaving your Kafka broker open to anyone who happens to stumble upon it does not qualify as “doing privacy right.” The Kafka broker operated without authentication or encryption, meaning anyone could listen in, collect logs, or potentially alter data if write access existed. This demonstrates a fundamental misconfiguration that puts both users and the platform at risk.

The Kafka instance contained over 621,000 log entries from the last 13 days, belonging to Huddle01, including:

• Usernames (sometimes real names)
• Email addresses
• Crypto wallet addresses (Huddle01 supports many wallets across blockchains like Bitcoin and Ethereum)
• Detailed activity data, such as which users joined specific calls, participants in each call, country, time, date, and duration
• Other identifiers

The app is popular among cryptocurrency users, but in this case the open Kafka instance could have deanonymized their wallets by tying their crypto wallets to usernames and email addresses. Which also paints a target on their back as potentially high-value targets.

It also makes users more vulnerable to social engineering since attackers can craft credible emails or messages using real names and meeting data.

And hold on for the worst part. Cybernews states it responsibly disclosed the data leak to the company behind Huddle01…

“However, it did not respond to the initial disclosure and subsequent attempts. After one month, the exposed server remained accessible. It’s unclear how many other third parties might have accessed the data.”

Security tips for affected users

Knowing that the exposed information goes back about two weeks doesn’t help much, since anyone with access could have set up a data collector, listening in on the real-time data streaming and processing going on.

So, any Huddle01 users should:

• Change passwords on accounts linked to the exposed email or username, and use strong, unique passwords for each site.
• Set up two-factor authentication (2FA) wherever possible to prevent unauthorized access.
• Monitor inboxes for suspicious messages. Be extra cautious of emails or texts asking for crypto transactions or sensitive information, as targeted phishing is a possibility. Be especially wary of social engineering attempts that reference details from meeting logs, such as who you spoke to or when meetings occurred.
• Stay updated on official statements from Huddle01 or news coverage, as they may release more guidance later.

Pro tip: Did you know that you can submit suspicious messages like these to Malwarebytes Scam Guard, which instantly flags known scams?

We don’t just report on scams—we help detect them

Cybersecurity risks should never spread beyond a headline. If something looks dodgy to you, check if it’s a scam using Malwarebytes Scam Guard, a feature of our mobile protection products. Submit a screenshot, paste suspicious content, or share a text or phone number, and we’ll tell you if it’s a scam or legit. Download Malwarebytes Mobile Security for iOS or Android and try it today!