• The US Treasury's FinCEN arm asks financial institutions to file scam center suspicious activity reports under a new keyword
• FinCEN says that it flagged financial moves across 33,904 different filings, but actual losses might differ significantly, as the approach is prone to double-counting transactions
• Only 1,300 institutions filed reports, with 10,082 (29.7%) of them centering around exploitation of the elderly by scammers

The US Treasury's Financial Crimes Enforcement Network has issuespublished an alert and a companion data analysis telling banks, credit unions, digital asset exchanges, and securities firms that it needs sharper reporting on the overseas scam centers that it says target Americans at an industrial scale.

FinCEN puts total damages, per its reporting mechanism, at roughly $12.7 billion linked to suspected digital asset investment scams between September 2023 and the end of 2025.

The number comes from adding up the dollar values of 33,904 Bank Secrecy Act filings that referenced the keyword from its 2023 "pig butchering" alert, which may be overstated, as it includes both attempted and successful transactions as well as both inbound and outbound reports, often of the same transactions, causing significant overlap.

FinCEN's new requirement is a keyword

While the $12.7 billion is a measure of what institutions have flagged and is prone to double-counting and errors, victim losses, a significant chunk of which go unreported, may be considerably higher.

This has prompted Gene Lange, who effectively works as the Under Secretary for Terrorism and Financial Intelligence, to call these scams "one of the most significant fraud threats facing Americans today."

It has also prompted a new suspicious activity report keyword: "FIN-2026-SCAMCENTERS," which institutions are expected to use to indicate that a scam center is potentially involved.

FinCEN also wants chat logs, scammer phone numbers, social media handles, wallet addresses, transaction hashes, and the URLs victims were told to deposit into, filed in the structured cyber indicator fields rather than left out.

Institutions to volunteer more information to stop scams

The move is part of its push to have institutions volunteer more information under the US Patriot Act, as it aims to confront what is a growing intelligence problem: scammers tend to route victims through several institutions in sequence; most filers see only one slice of a scam's lifecycle and often have difficulty tracing it all the way.

For example, a crypto exchange might see a customer buying USDT and sending it off-platform, a bank might see a wire to that exchange, and a brokerage might see a retirement account liquidated. None of them would have the bigger picture of what essentially happened or what triggered the transaction.

Combined, with properly linked information, it makes it much easier to identify a potential scam; separately, these incidents can inflate the number of reports, which often aren't linked, and investigations can lack insight into the origin or final destination of the funds.

FinCEN's Rapid Response Program has interdicted $1.8 billion and recovered just over $1 billion for 5,790 US victims since 2015, which, against the flagged totals, is a very limited recovery at best compared with the actual funds at stake. This highlights a larger fundamental problem: institutions mostly detect these schemes after the money is gone, and reporting them correctly and thoroughly may yield limited dividends at best against an industry that has morphed, relatively unchecked, into a multi-billion-dollar juggernaut.

• A pseudonymous GitHub account has published what it calls a reconstruction of Stuxnet malware that purportedly took out a fifth of Iran's centrifuges before being discovered
• The original source has never surfaced, and the 'Stuxnet' moniker that the account uses comes from Symantec's coining of the name weeks after it was discovered
• The code remains unverified and untested, with some users indicating that it is likely an AI-generated replication of the original binaries' behavior, which has been widely documented

A GitHub account has published what it describes as a faithful reconstruction of Stuxnet, the worm that sabotaged Iranian uranium-enrichment centrifuges.

It also became the first piece of software widely accepted as having caused physical destruction in the real world, highlighting how malicious software can often do much more harm to people and hardware than just stealing or manipulating data.

The repository surfaced via a Show HN submission, which mainstream media outlets then picked up, highlighting the as-yet-unknown researcher's reverse-engineered code.

Not the first or last Stuxnet repository

The original Stuxnet source, written by whoever built it, has never been leaked and did not leak last week. The repository's README explicitly states that it is a reconstruction assembled from decompiled binaries, and those binaries have been in public circulation since Belarusian firm VirusBlokAda pulled samples from an Iranian customer's machines in June 2010.

Everything that followed, including Symantec's W32.Stuxnet Dossier and Ralph Langner's To Kill a Centrifuge, was built on those samples and how they reacted in test environments.

This is also not the first time readable C-language code aiming to replicate Stuxnet has been published online. Malware researcher Amr Thabet published a decompilation of the MRxNet rootkit carrying a 2010 to 2011 copyright notice. Christian Roggia followed with a dropper decompilation called open-myrtus, copyrighted 2012 to 2014, which has since been forked into a long chain of repositories.

The irony is that if such code were a faithful replication of the infamous malware, it would not have mentioned "Stuxnet" in multiple places, including registry keys. This is because the moniker is not one the developers likely used, but one Symantec switched to from its original 'W32.Temphid' identifier.

The thread where it was first brought to attention by a user called Sadpainy also has mixed views from developers, many of whom have branded it 'AI slop' or a 'fake' that relies on a mixture of already existing repositories, even as the about page on its GitHub repo states that it was reproduced by the researcher for educational purposes and is designed to only work on Windows XP and Windows 7.

For those looking to test it, a virtual machine might be their best bet, especially given Stuxnet's ability to physically damage hardware, but it is also a stark reminder of what a rogue AI agent could do if left unchecked without specific instructions or safeguards.

• UK NCSC, FBI, and Dutch AIVD warn Iran is using Chosen Brick malware against dissidents and journalists
• Malware steals files, captures audio, grabs WhatsApp/Telegram data, and can wipe systems entirely
• Operatives rely on social engineering; agencies urge awareness, MFA, updates, and endpoint monitoring

Iranian hackers are targeting “enemies of the state”, both local and foreign, with advanced malware capable of spying on the victims and stealing their sensitive files, experts have warned.

This is according to a new security advisory, published jointly by the UK National Cyber Security Centre, the FBI, and the Netherlands’ General Intelligence and Security Service (AIVD), which noted how Iranian operatives would first do extensive research into their victims - dissidents, activists, and journalists - deemed a risk to the regime.

After learning as much about their targets as possible, they reach out via social media, either as someone the victims know, or as technical support for the platform they’re currently using, engaging in extended conversation until the victim lowers their guard. At one point, the attackers will try to share a piece of malware with the victims, tracked as Chosen Brick.

A thousand victims

This malware, designed primarily for the Windows platform, has a long list of capabilities, including enumerating running processes and system information, capturing screen content, enabling the microphone to capture audio content, capturing a copy of Telegram and WhatsApp data from web browsers, downloading additional files and malware, deleting files, stealing email content, and ultimately - wiping the entire computer system. The operatives communicate with the malware using Telegram, it was said.

“Iran almost certainly uses cyber activity to support the repression of individuals who are seen as a threat to the regime, such as dissidents, activists and journalists,” the three agencies said in the report. “In some cases, the Iranian intelligence services have plotted to kidnap or conduct lethal operations against individuals internationally, who they perceive as enemies of the regime.”

In the advisory, the three agencies said the best defense is to simply be more aware of social engineering. However, there are also a few technical mitigations that can help, including following NCSC advice on staying safe online, keeping all devices up-to-date (ideally through automatic updates), using antivirus software, and not disabling smart screen warnings on file downloads.

Finally, it would be wise to enable phishing-resistant MFA, make sure devices are managed with appropriate controls, turn on email scanning, deploy endpoint and network monitoring, and conduct a search for the IoCs.

Via The Register

• CenterPoint Energy confirmed a cyberattack via exposed API, with customer data stolen
• Threat actor claims 7.49M files including IDs, SSNs, billing data, and transaction records
• Operations unaffected; investigation ongoing, regulators notified, customers to be informed

CenterPoint Energy has confirmed suffering a cyberattack and data theft, days after a criminal advertised the stolen files on an underground hacking forum.

CenterPoint Energy is a large US energy utility company that delivers electricity and natural gas to homes and businesses. It employs roughly 8,800 people and operates around $48.3 billion in assets, as of June 2026.

Recently, a threat actor posted a new thread on a dark web forum, saying they stole 7.49 million CenterPoint files from a poorly secured API, The Register reports. They said that the data included customer names and contact details, billing data, move-in dates, driver’s license information, and the last four digits of Social Security numbers (SSN).

Incurring expenses

No independent investigators have confirmed these claims just yet, and the company said it was investigating the matter. In a new 8-K document filed with the US Securities and Exchange Commission (SEC) on September 14, the company said that it “became aware of an online post by a third party claiming to have obtained a data set containing certain of the company’s customer information.”

It activated its incident response protocols and kicked off an investigation with the help of third-party cybersecurity experts.

“While the investigation remains ongoing, the company has determined that an unauthorized third party obtained personal information relating to a portion of the company’s customers through one of the company’s external facing systems,” the filing reads. “The company is continuing to work with third-party experts to determine the scope of customers and personal information affected by the incident and intends to notify affected customers and regulatory authorities as required by applicable law.” The police and regulatory agencies have been notified.

While the attack did not impact CenterPoint’s operations, which continue as usual, it did incur certain expenses, the company concluded. It stressed that it will likely incur even more expenses as the investigation continues.

Via The Register

• Defiant warned of active exploitation of WooCommerce Wholesale Lead Capture Plugin flaw (CVE‑2026‑27540)
• Critical unauthenticated file‑upload bug lets attackers deploy PHP webshells for site takeover
• Patch released in Feb 2026 (v2.0.3.2); Wordfence blocked 100,000+ attacks, users urged to update and check uploads

A critical vulnerability in a popular WooCommerce plugin is being actively exploited to upload malware and possibly take over entire websites, security experts have warned.

The plugin in question is called Wholesale Lead Capture Plugin for WooCommerce. It adds a dedicated registration and onboarding system for wholesale and B2B customers, letting businesses collect company and other custom information, review applications, assign wholesale user roles, and automate registration and onboarding emails.

It is a premium plugin that costs between $99 and $600 and which, according to the Wordpress store page, has more than 20,000 active installations.

A thousand victims

The plugin was vulnerable to an unauthenticated arbitrary file-upload flaw which, as the name suggests, allows unauthenticated parties to upload arbitrary files, including PHP webshells and executable code that can result in full site takeover. It is tracked as CVE-2026-27540 and carries a severity score of 9.0/10 (critical).

Versions 2.0.3.1 and older were said to be affected. Version 2.0.3.2, released on February 20, was said to have addressed the bug, meaning the patch has been available for almost half a year. However, WordPress security outfit Defiant said its Wordfence web application firewall blocked more than 100,000 attacks, and Wordfence added that it observed two attack spikes - one between June 4 and 17, and another one between July 1 and August 30.

In these incidents, the attackers were mostly uploading reconnaissance webshells, possibly mapping out the landscape before deploying more serious malware.

“The uploaded shell.php is a PHP webshell that reports host details and provides a browser-based upload form for writing additional malicious files to the site,” the researchers said.

If you are running the plugin, it is advised you update it to the newest version as soon as possible, and check upload directories for unexpected, or recently created, PHP files.

Via BleepingComputer

• Elastic Security Labs uncovered REF9334, a Brazilian banking malware campaign active since May 2025
• Malware “Kremlin” deploys fake docs and malicious Chrome/Edge extensions to steal banking data
• 1,515 infections found, 98% in Brazil

Security researchers from Elastic Security Labs have discovered a new Brazilian banking malware campaign that uses browser extensions to compromise users and steal sensitive information.

In an in-depth report published earlier this week, the researchers said the campaign has been active since at least May 2025. Dubbed REF9334, the campaign uses fake banking, invoice, and business documents, to trick victims into installing malware which, in turn, deploys a malicious extension in Chrome and Edge browsers.

The researchers named the malware “Kremlin”, and say it can steal browser credentials, cookies, session information, monitor browser activity, grab screenshots, and steal information from websites that the victims visit. But the goal of the campaign is primarily to target Brazilian bank users.

A thousand victims

The malware really makes an effort to hide and persist in the target environment. For example, it first checks to see if it’s in a sandbox and if so - it simply won’t run. If instead it determines that it’s running on a real user’s computer, it will deploy an extension with the name “AVSync System Inc.” in an attempt to trick the victim into thinking they have an antivirus addon running in the browser.

It also doesn’t use a fixed C2 server, but rather stores the information on the Ethereum blockchain, since it’s a lot harder to disrupt the communication between the operators and the infected machines that way.

During their investigation, Elastic researchers were able to take control of a domain that the malware used and discovered that it had infected 1,515 systems. Almost all of them (98%) were located in Brazil. They were also able to register the network canary domain and point it to their webhost, which resulted in the loader assuming it was in a sandbox. This also meant “the infections have not moved past the initial access”, Elastic explained.

The full list of indicators of compromise can be found on this link.

Via The Hacker News

• 153 million US driving licences have been leaked on a Russian cybercrime platform
• Among those apaprently discovered in the stolen data is US Secretary of Defense Pete Hegseth
• The FBI is now investigating the leak, which has been traced to an identity verification company

A data leak of 153 million US drivers licenses is said to have been shared on a Russian cybercrime forum, with US Secretary of Defense Pete Hegseth among those leaked prompting an FBI investigation.

Security researcher Brian Krebs identified the leak – which included his own data – as originating from a hack of an identity verification service. Louisiana-based IDScan provided ID verification for various well-known companies, including FedEx and Hertz car hire.

The data was shared on a Russian forum called Exploit, a long-established online community of cybercriminals. Following news of the leak, the identity theft service “Nexus” has apparently scrubbed its existence from the Dark Web.

Driving licenses and more

It wasn’t just US driving licenses that were found in the archive of recently-collected personal data. Krebs’ investigation found that Nexus claimed other types of data, and found a further 1.1 million driving licenses from Canada.

Other identity documentation alleged to be in the leak include 10 million identification cards, three million travel documents and international IDs, and 579,000 medical cards. The data was available to browse, notes Krebs, with Nexus providing details: “Records are available to preview before purchase with pertinent information redacted. Customer photos are displayed if available.”

The leak has a personal dimension for Krebs. Not only was his driving license in the collection, so was that of his mother. It has proved to be a useful coincidence, one that has enabled the security and privacy researcher to establish how the data was sourced by Nexus.

Both licenses were used for a car hire, which Krebs traced to Hertz. Others affected by the leak had also used the service, which has used New Orleans-based IDscan for identity verification.

The company, which claims to perform 21 million verifications a month, is yet to issue a statement on the matter. Its marketing and operations leader, Jillian Kossman, told the journalist: “At this point I’m not able to share any additional information, but the updates you have provided have been welcome, and helpful to our team’s investigation.”

Where is the data?

Krebs reports that he was alerted to the data on August 31, 2026, little over a week ago. Regular checking noted that the data was still being added to, increasing by “nearly 400,000” records prior to publishing his investigation on September 1.

Since then, however, it seems that Nexus has vanished, along with the data. But how widely was the data downloaded before that happened? While the FBI investigates, it falls on the American and Canadian public to be extra vigilant and wary of identity fraud.

• Huntress warns fake GTA 6 ISOs spreading malware ahead of official Nov 19, 2026 release
• Packages deliver RATs, infostealers, and Chaos ransomware disguised as game installers
• Users urged to avoid pirated downloads; infected systems should be reimaged and credentials reset

Grand Theft Auto 6 (GTA 6), one of the most anticipated computer games of all time, is due to launch on November 19, 2026 - so everything found online before that is almost certainly malware.

This is the warning sent out by security researchers Huntress, which said it recently saw one variant deploying all sorts of filth, from remote access trojans (RAT), across infostealers, to full-blown ransomware encryptors.

The story behind GTA 6

GTA 6 is a third-person action-adventure and open-world gam developed by Rockstar Games, which egan preliminary work more than a decade ago, and has allegedly spent more than a billion dollars during that decade and a half, which would make it, by far, the most expensive game ever developed.

It is also a sequel to GTA 5, a game that was sold in more than 230 million copies around the world, which made it one of the most popular games of all time.

In late August 2026, someone going by the alias LEEK claimed to have obtained a working, playable build of the game, roughly six months before the official release. They posted numerous screenshots and videos of them playing the game, even writing their nickname onto a wall, using a rifle to prove the authenticity. Truth be told, even without this, people would have still probably searched for, and downloaded, fake installers. This only made the problem worse.

And a major problem, it definitely is. Huntress is now saying it is seeing websites ranking relatively high (through SEO poisoning), offering for download an ISO of the game. Similar ISOs are circulating around different gaming forums, torrenting sites, and social media channels.

An ISO file is a digital copy of an entire disc, such as a CD, DVD, or Blu-ray, stored as one file. Users can download an ISO (which can weigh tens or hundreds of gigabytes), mount it into a virtual drive, and the computer will treat it like a physical disc.

Huntress found some ISOs weighing more than 100GB which was done just to make it seem authentic - the actual malware was “a factor smaller”. The ISOs they analyzed contained multiple variants, including NJRAT and CDRAT (two remote access trojans), Mercurial Grabber (an infostealer grabbing Roblox Studio cookies, Minecraft session data, Discord tokens, Chrome passwords and cookies, system information, IP addresses, and geolocation, Windows product keys, and screenshots), and Chaos Ransomware. The package also contained a legitimate, regular browser - “Just for kicks,” apparently.

A plausible failure

During the “game” installation, the installer displays a message in Russian, saying the product is unlicensed and that there is a good chance it won’t run. If that happens, the user is told to reach out to a specific Gmail address to receive an “updated crack”. After the installation completes, the victim will get that exact popup, saying “License not found”. All of this is just a smoke screen, to make the victim not suspicious as to why the “game” won’t run.

There is no way of telling how many people fell for the ruse and installed malware on their devices. Huntress says that generally speaking, trying to download cracked and pirated software is a bad idea, which is even worse if the game in question has not yet been released.

“This is fertile ground for scams and threat actors attempting to take advantage of the impatient and overeager,” the researchers said.

The good news is that all of the malware being distributed there is not new. Some of the variants contained in the ISOs are several years old, meaning most antivirus programs, such as Windows Defender for example, can easily detect it and stop it from compromising the system.

Those that do end up infected should disconnect the machine from the internet, reset all passwords, enable 2FA wherever possible, and do a complete reimage of the compromised system.

• Proofpoint detailed BlueMoon, an exploit kit chaining two Chromium flaws and one Windows bug
• Four groups, including China‑aligned TA412, used it loudly against NGOs, aerospace, and manufacturing targets
• Exploits were “patch‑gap” zero‑days; all flaws now patched

Four hacking groups, including some tied to the Chinese government, were seen using the exact same exploit kit in a span of a week, suggesting a certain “fear of missing out” among the criminals, experts have warned.

Security researchers Proofpoint have detailed BlueMoon, an exploit kit that leverages three vulnerabilities: two in Chromium, and one in older versions of Windows: Windows 10 (October 2018 Update), Windows Server 2019, Windows 10 2004, WIndows Server 2022, and the initial release of Windows 11.

The kit was first seen being used on August 28 2026, by a threat actor tracked as TA412, a China-aligned state-sponsored threat actor that was observed in the past targeting businesses using Microsoft SharePoint. TA412, also known as Violet Typhoon, used BlueMoon to “repeatedly target a small number of non-governmental organizations (NGOs), mining companies, and physical commodity trading firms in the United States.”

Soon after that, three others were spotted: UNK_LateNight, another China-aligned espionage group which targeted multiple US aerospace companies, UNK_DoubleCheck, a group going after a Vietnamese manufacturing entity, and UNK_QuietRacket, a threat actor taking aim at organizations across Singapore and Indonesia.

What all of these groups have in common is the fact that they did not try particularly hard to hide their activity. This is rather unusual, since a stealthy approach usually guarantees that a vulnerability can be exploited for longer.

Front running the Chromium supply chain train

BlueMoon takes advantage of three flaws, two in Chromium and one in older Windows variants. The Chromium vulnerabilities were found in V8, the JavaScript engine that the browsers use to allow applications to run efficiently. The first one is a “type confusion bug”, tracked as CVE-2026-85046, and assigned a severity score of 8.8/10 (high). The latter is a “sandbox escape” flaw, for which Google did not assign a CVE or a severity score.

The Windows bug, on the other hand, is tracked as CVE-2026-85880, and was given a severity score of 7.8/10 (high). It is described as a “heap-based buffer overflow” vulnerability in Windows Advanced Local Procedure Call, allowing malicious actors who can already execute code inside a low-privilege AppContainer to escape the sandbox and elevate privileges to SYSTEM. No additional user interaction is required.

Proofpoint has a theory as to why the crooks decided to be loud, instead of flying under the radar. Apparently, there is a very short window of opportunity between Google patching a Chromium vulnerability, and it being deployed to a browser such as Edge, or Brave, This window of opportunity also allows crooks to see how Google fixed a flaw, reverse-engineer it, and deploy an exploit before the browser is patched. This would mean that there is no time to be stealthy.

No time to hide

“Both V8 vulnerabilities were ‘patch-gap’ zero-days at the time of the observed activity,” Proofpoint said. “In other words, while they were known vulnerabilities already fixed in public upstream Chromium source code, they remained unpatched in the latest stable releases of Chrome and Chromium-based browsers available to the public. It is likely that the exploit kit developer used these publicly available Chromium patches to weaponize the browser exploit chain.”

Another important factor is Artificial Intelligence. It would seem that AI has made flaw detection significantly faster, reducing the barrier to entry and making threat actors move more loudly.

“A fully weaponized Chrome exploit chain has historically been a high-value, rare capability. BlueMoon was developed, deployed rapidly, and shared across multiple threat actors within days in a manner that had high detection signals. This may reflect a reduced cost and barrier to entry for this class of capability, as AI agents increasingly enable threat actor exploit development,” Proofpoint stressed.

All three flaws have since been patched, so make sure you’re running the latest version of both the OS and the Chromium browser.

Via Ars Technica

• Calif researchers found a zero‑click WeChat VoIP flaw enabling account takeover via calls
• “WeWorm” spreads through ringing calls; victims need not answer to be compromised
• Tencent patched in Android 8.0.77 and iOS 8.0.76; no exploitation seen in the wild

Security researchers have found a flaw in WeChat which allows malicious actors to take over people’s accounts on both Android and iOS devices - but what makes this flaw stand out is the fact that it’s a zero-click bug - victims need not do a thing to be compromised.

WeChat is a “super-app”, allegedly used by roughly 1.4 billion people, and is especially popular in China. It started as a communications app, letting users send messages, and make voice and video calls, and has evolved to function as a social network, allowing users to share photos and videos, as well as a payment app through which users can transfer money, pay for things, order food, book taxis, and even access government and business services.

Security researchers from Calif have now disclosed finding a ‘memory corruption’ issue in WeChat's VoIP stack. For now, they decided not to share the technical details, and to instead demonstrate the flaw “at an upcoming conference.” To that end, they built a worm called WeWorm, capable of taking over target WeChat accounts and spread through phone calls made via the app.

A phone call would suffice

In practice, it works remarkably simple: an attacker uses WeChat to call a person they have in their contacts list (this is a prerequisite). They can use both an Android and an iOS device, and can call anyone, regardless of the model or the OS they’re using. As soon as the phone starts ringing, WeWorm gets to work, “worming” its way into the victim’s device.

The victim does not even need to answer the phone - having it ring is enough. If they answer, they’ll hear nothing but silence, yet the worm will continue operating. If they decline the call, the attack stops, but this is hardly a mitigation - the attacker can simply call again while the victim is asleep (or otherwise away from their device).

Within a few seconds, the attacker will have access to the victim’s WeChat account, including their messages, contacts list, and virtually anything else found in the app. What makes this bug particularly worrisome on the surface is the fact that WeChat can be used to transfer money and pay for things, but WeChat Pay has additional authentication and risk controls designed to prevent that from happening.

The good news is that there is no evidence of this flaw being exploited in the wild. The bad news is that this is not the first zero-click flaw found in modern-day smartphones, and most likely will not be the last one.

Tencent's response

Calif said it responsibly disclosed its findings to WeChat’s parent company Tencent, who came back with a patch. Versions 8.0.77 for Android and 8.0.76 for iOS have apparently solved the problem, although

Tencent did not list any details in its patch notes, simply saying the version brought “bug fixes”, but in a statement shared with The Hacker News, it said the exploit has been “mitigated for all users”, and that it was applied server-side - users need not install anything, aside from the patch.

It’s also worth mentioning that WeChat has apps for HarmonyOS, Windows, Mac, and Linux. However, it would appear that Calif did not test those, and Tencent did not include them in its patch. The researchers did say that they would be looking into this same flaw in other products, too:

“This specific WeChat bug is one instance of the many unconventional attack surfaces that are present across many messaging apps,” they said. “We're conducting more of this research across other apps and attack surfaces, while working with app developers on attack surface reduction. This may take an industry-wide effort, since some of it depends on the platform owners. Once that work is further along, we'll share our progress, including the technical details of this WeChat bug.”

• CISA, NSA, FBI warn Chinese AI firms of industrial‑scale knowledge distillation
• Companies like DeepSeek, Moonshot, Alibaba allegedly extracted billions of tokens from US frontier models
• Advisory urges detection of malicious prompts, deceptive responses to distillation, and cross‑provider intelligence sharing

Chinese AI companies’ core development strategy is to steal proprietary functionalities and capabilities from their US counterparts, law enforcement agencies have warned.

The US Cybersecurity and Infrastructure Security Agency (CISA) has published a new security advisory, drafted jointly with the National Security Agency (NSA) and the Federal Bureau of Investigation (FBI), warning American AI companies about an ongoing “aggressive, malicious, and targeted distillation activities at an industrial scale,” and sharing recommended mitigation steps.

Knowledge distillation

IBM defines knowledge distillation as a “machine learning technique that aims to transfer the learnings of a large pre-trained model, the ‘teacher model,’ to a smaller ‘student model’.” It is used in deep learning as a form of model compression and knowledge transfer, it added, particularly for massive deep neural networks.

So, knowledge distillation is not illegal or malicious, per se. Its goal is to train a more compact model to mimic a larger, more complex one. In the security advisory, the agencies stress it is “recognized as a legitimate and useful technique in AI research,” but add that China-based AI companies are using it in ill will.

In other words, the agencies claim that instead of spending months and millions developing new capabilities for their models, the Chinese are simply sending huge numbers of carefully designed questions to US models and extracting the answers.

Which companies are engaged in knowledge distillation?

Apparently, all companies worth anything. DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI all allegedly “extracted billions of tokens across millions of exchanges/requests from US frontier AI models, including variants of Claude, GPT, Gemini, and Grok, since at least late 2024.” CISA also stressed that this was likely done with the awareness of the Chinese government. It hasn’t outright said, “with its blessing”, although it could be read between the lines.

The advisory shares a thorough list of all the models that were being trained, as well as all the models being taken advantage of.

On the Chinese side, they include DeepSeek R1 and V3 models, Moonshot’s Kimi-K2 and Kimi-K3 models, and MiniMax’s M2 model. On the US side, they start with earlier models such as GPT-4, Claude 3.7, and Gemini 2.5 Flash Preview, all the way to Claude Fable 5, GPT-5, and similar.

When done in good faith, knowledge distillation is not illegal. However, the report says the companies routed the requests through multiple accounts, different API access points, multiple cloud providers, third-party AI aggregators, proxy services and “transfer stations”, as well as premium subscriptions shared between developers, all in an attempt to work around defenders trying to disrupt the process.

“This represents systematic extraction of proprietary functionalities and capabilities threatening U.S. technological leadership. Addressing industrial-scale distillation merits a coordinated response across the AI ecosystem, including effective information-sharing, spanning the U.S. Government, private industry, and allied nations,” the agencies concluded.

What US companies should be doing

To defend their intellectual property (and thus remain ahead of Chinese competing models) US AI companies should implement comprehensive detection and mitigation, the agencies said. That means hunting for anomalous and malicious prompts, accounts, networks, and behaviors. Furthermore, they should monitor subscription-to-usage ratios, immediate maximum usage from new accounts, and enterprise-scale throughput patterns.

The second step is to “deploy targeted response changes”: “Subtly alter responses for suspected malicious distillation attempts to attenuate the payoffs to companies conducting industrial-scale distillation campaigns.” In other words, AI companies should make sure their products lie when they spot they were being distilled for knowledge.

Finally, US AI firms should set up cross-organization intelligence sharing, correlating activity across model providers, cloud platforms, and API aggregators.

• Check Point Research exposed coerced insider flaw in ChatGPT’s agent architecture
• Containers shared metadata via internal service, enabling cross‑account prompt injection and data theft
• OpenAI closed the path, but CPR warns similar risks may exist in other AI platforms

ChatGPT’s AI agents were allowed to pull sensitive data from one account shared with an entirely different account because, colloquially speaking, all agents used to walk down the same hallways, experts have warned.

A new report from security experts Check Point Research (CPR) dubbed the flaw “coerced insider”, since it revolves around persuading the agent instead of abusing a vulnerability.

Coerced insider

When an AI agent is given a task that needs code execution, it handles that task in an isolated container which also sometimes needs to install software. To enable that, without giving containers direct internet access (which would be too risky), OpenAI routes those packages through an internal JFrog Artifactory instance. As a separate security contingency, containers from different accounts cannot talk between themselves.

However - they can reach the same internal service (our proverbial hallways), which exposes an item management feature that lets the containers attach text or binary properties to a repository item. As a result, any container can read back the properties written by any other container.

“Check Point Research confirmed the isolation gap directly: a property written from one account’s container was fully readable from a different account’s container moments later, with data too large for one property simply split into chunks and reassembled on the other end,” the researchers explained.

“The package delivery metadata effectively became a shared clipboard between containers that were supposed to be walled off from one another.”

From there, the exploit turns into your usual, off-the-shelf prompt injection. The only difference is that the malicious prompt is not delivered directly to the victim, but rather left in the hallways, and the results are not shared with the attackers directly, but rather left in those same proverbial hallways, too.

The attacker then delivers a prompt or a shared conversation that instructs the agent to check the same storage during its next ordinary reply. The agent checks, sees the malicious instructions, executes them, all the while replying to the victim’s question in the usual manner. The victim is oblivious to the fact that data theft is taking place in the background.

What kind of information can be stolen?

But this is just half of the equation. How destructive this attack ends up being still depends on the amount of data being shared with the victim agent.

The bare minimum is the information shared while chatting to the agent. It then grows with every connected app: Gmail, Google Drive, Microsoft Teams, GitHub, and similar. “In Check Point Research’s demonstration, ChatGPT retrieved the victim’s email data through their connected Gmail account and delivered it to the attacker’s session, all within a single ordinary turn,” CPR stressed.

The good news is that you’ll likely never be exposed this way, at least not via ChatGPT. CPR says it disclosed the findings to OpenAI, who then confirmed that the specific internal Artifactory instance identified in the research has been commissioned. In other words, the hallways attack path has been closed.

The bad news is that this doesn’t automatically mean everyone’s safe. This particular path might be closed, but the architectural pattern behind the flaw could be present in other platforms, CPR warns.

“Any AI assistant that operates inside an organization’s trust boundary, holding credentials, running code, and reaching connected services, can become what Check Point Research calls a coerced insider,” the report states. “The model itself does not need to be malicious. It only needs to be persuaded, through text it was never meant to trust, to use access that was granted for entirely legitimate reasons.”

Going forward, businesses are advised to learn which AI tools their employees are using, and what those tools are connected to. Then, they should govern what AI tools and agents are allowed to do, treating all of their actions (not just output) as something that needs to be monitored.

• OpenAI expands Daybreak with Blue and Red tiers for defensive cyber work
• New GPT‑5.6‑Cyber model offers high compliance for authorized vulnerability research
• Access remains restricted due to dual‑use risks and reduced safeguard operation

OpenAI has announced two new tiers for its Daybreak dedicated cybersecurity project, each offering a different model with different levels of compliance. It also used the opportunity to introduce a new security-focused AI model, as well.

Hackers and criminals are increasingly abusing AI to improve and speed up the creation of phishing emails and malicious code, and with the introduction of AI agents, they’ve also used it to automate entire attack processes. The AI community responded by placing strong guardrails, making sure their models do not comply with requests to build malware, or hack other companies.

These guardrails ended up being a two-edged sword, because as they slowed down attackers, they also slowed down the defenders.

What is Daybreak?

To give the cybersecurity community the upper edge, companies like OpenAI started creating dedicated cybersecurity initiatives that provide a vetted list of companies state-of-the-art models, free of guardrails. The company also provided them with pre-trained AI agents, as well as access to a pool of shared knowledge.

Initially launched in June 2026, Daybreak originally included GPT.5-5-Cyber (a model optimized for security work), Codex Security (an agent that can analyze codebases, identify vulnerabilities, validate findings, and help develop patches), Patch the Planet (an initiative with Trail of Bits to find and fix vulnerabilities in open-source software), Daybreak Cyber Partner Program (lets approved cybersecurity companies such as Cloudflare or Cisco integrate OpenAI's cyber capabilities into their own products and services), and Trusted Access for Cyber (the governance/access system for organizations doing authorized cybersecurity work with these capabilities).

Now, OpenAI has expanded Daybreak with two access tiers, Daybreak Blue, and Daybreak Red.

The company says Daybreak Blue is “the recommended starting point for most defenders, supporting vulnerability discovery, secure code review, malware analysis, incident response, and patch validation. Companies opting for this tier can expect access to frontier general-purpose models, including GPT‑5.6 Sol, whose safeguards have been tailored to authorized defensive security work.

Daybreak Red, on the other hand, provides access to OpenAI’s “purpose-trained cybersecurity models for authorized vulnerability research, exploit validation, and security testing.” This tier offers the brand new GPT‑5.6‑Cyber, built on GPT‑5.6 Sol and trained to improve capabilities on several specialized cybersecurity tasks such as finding zero-day vulnerabilities and developing exploit chains.

This model is also more compliant and less likely to refuse certain higher-risk, dual-use cyber tasks.

Complying with "dangerous" requests

Request compliance is the name of the game here. OpenAI says the new model addresses feedback from security researchers who “encountered persistent refusals with the earlier model.” General-purpose GPT-5.6 Sol, for example, will comply with just 1.5% of the requests usually given by cyber-defenders working on codebase analysis or vulnerability identification. This percentage increases to 2.0% with Daybreak Blue access.

GPT-5.6-Cyber, on the other hand, completes 95.0% of requests, OpenAI says, up from 57.3% of the previous model, GPT-5.5-Cyber. We weren’t able to independently verify these claims, though.

While it doesn’t outright say it, OpenAI considers these models relatively dangerous to use, which is why they’re locked behind the Daybreak Cyber Partner Program. However, the program is now expanding, allowing these companies to embed the models behind their own products, managed services, or cybersecurity engagements, and offer them to clients of their own.

Those who wish to be a part of the program directly can do so by applying to join online now.

“Models running with reduced safeguards carry risks beyond standard model usage, whether from misuse or misalignment. Despite these risks, we believe that democratizing access to frontier intelligence for defenders is crucial to accelerating and automating cyber defense,” OpenAI said.

“Daybreak Blue and Daybreak Red access are available for approved individuals⁠ and organizations conducting authorized work. We control access through identity verification, account security, monitoring, approved-use restrictions, and legal attestations.”

• CEVA Logistics hack disrupted European warehouses and exposed some customer data
• Retailers and Valve reported compromised delivery information and service delays
• Clients warned of targeted scams while awaiting fuller incident details from CEVA

CEVA Logistics, one of the biggest shipping and logistics companies in the world, has suffered a major cyberattack, the effects of which are trickling down to many of its clients.

The details of the hack itself, however, are not yet publicly available and what little information is out there came from the affected clients themselves.

CEVA has not yet issued an official statement, or filed a report with the regulators, but confirmed to TechCrunch that the attack most likely started on July 29, 2026, and affected at least eight warehouses across Europe.

Technical details missing, affected customers step forward

CEVA Logistics is a global logistics and supply-chain company and a wholly owned subsidiary of CMA CGM, a French shipping giant. It provides freight forwarding, contract logistics, warehousing and transportation services to thousands of customers, including major companies in the consumer and retail, automotive, industrial and aerospace sectors.

The company operates in more than 170 countries around the world and last year it generated $18.3 billion in gross revenue, so it is a major player and a key target for attackers.

The details about the incident itself are scarce right now. We don’t know how the crooks broke in (via a successful social engineering attack, or by abusing a software vulnerability, for example), how much data they stole, or if they demanded a ransom payment in exchange for deleting the stolen goods. From one of the victims, though, we have learned that some data was most likely compromised.

When the effects of a cyberattack spill into the physical realm (as is the case here with eight affected warehouses) we can speculate the attack was either ransomware, or disruptive malware. Companies shut down parts of their IT infrastructure only when there is no other way to clear an infection or remove malicious outsiders.

In the meantime, a small number of CEVA’s customers confirmed suffering an attack and losing sensitive data.

Among them is Bol, a Dutch online retail company, which said the incident affected two systems used for processing orders from one of its fulfillment

“No bol systems were affected,” it said. “However, data of customers whose orders were processed via this location may have been viewed or copied.”

Bol also said restoring operations at one of CEVA’s locations was taking longer than anticipated. As a result, the assortment stored at the affected location was taken offline, and the products were unavailable for sale. Also, Bol is currently unable to receive goods from suppliers and sales partners at that location.

A similar announcement was given by De Bijenkorf, another Dutch luxury retailer, who said that order processing, returns, and refunds, might take longer, but stressed that its stores remained open. It also said that some customer data may have been compromised, including names, contact details, online orders data and, in some cases, VAT numbers. Payment information, bank account numbers (IBANs), credit card information, usernames, or passwords, were not compromised, it was confirmed.

Valve steps forward

Retail giants aside, PC gaming powerhouse Valve also notified its customers about the incident. It said CEVA ships Steam hardware to its European customers and as such, receives specific delivery-related information from Steam.

This information, which CEVA retains for up to 90 days after the order, was most likely compromised. It includes names, street addresses, phone numbers, email addresses, and the type and price of ordered products.

Valve warned its customers to expect fake messages, either via email, SMS, or phone, that might mention recent hardware orders.

“They may quote your address back to you to prove they're genuine. They may ask you to confirm a delivery, pay a small customs or redelivery fee, or sign in somewhere to “verify” your order. Treat all of them as fake,” Valve warned. The company also stressed that customer accounts are safe and that users need not do anything to secure them.

Other details are missing, not just for the general public, but for the affected CEVA clients, as well. Valve said it was “pressing” the company for the full scope of what was taken and how, and added that it is notifying the relevant authorities, as well.

A spokesperson of the Dutch data protection authority, Mark Schenkel, told TechCrunch the agency so far received 10 incident reports. Given the size of CEVA, it’s safe to assume there will be others.

Over the past month, it seems like every frontier model has broken free of its constraints and launched a devastating attack against one or more other companies.

One of OpenAI’s models escaped a testing sandbox and launched a very real attack against AI and machine learning company Hugging Face. Just days later, Anthropic revealed that multiple variants of its Claude model also escaped a sandbox that wasn’t properly sealed and began attacking the enterprise infrastructure of three companies.

Now, Meta has revealed that one of its models attacked another company’s infrastructure during testing. The accident has been pinned on a misconfiguration that allowed the model to access the internet. So why have so many incidents happened in such a short space of time?

Why are models escaping their sandbox?

In the cases of Anthropic and Meta, their models were being tested by a third party company called Irregular. Anthropic’s AI model was taking part in a "Capture the Flag" exercise, where the model’s raw offensive capabilities were tested without the usual safeguards. But the sandbox was left connected to the internet. A similar error to Meta’s own accidental escape.

During the OpenAI incident, the company was testing two versions of GPT‑5.6 Sol using the ExploitGym benchmark. Unfortunately, the AI models performed better than expected - chaining multiple attack vectors, stolen credentials, and zero-day vulnerabilities.

The main reason these models are escaping their testing environments is because they are designed to do exactly that. These AI models act like a massive team of highly-trained cybersecurity experts hunting for vulnerabilities and exploits. But what would take a team of humans days or weeks to accomplish can be done in hours, or even minutes, by these AI models.

It’s no wonder thousands of employees from AI firms are calling for a pause on the development of the technology, and Congress is considering an AI kill switch.

Expert perspectives on AI escapes:

OpenAI

• Nathaniel Jones VP, Security & AI Strategy, Darktrace:

What makes the OpenAI and Hugging Face incident important is that the models did not need malicious intent to cause harm. They were given the legitimate goal of solving a cybersecurity benchmark and found an unexpected route to the answers, escaping their test environment and compromising another organization in the process. From the models’ perspective, this appears to have been an effective solution to the task.

The AI's actions challenge the assumption that giving an agent a legitimate goal will produce legitimate behavior. As models become capable of pursuing objectives over longer periods, developers need to define not only what success looks like, but also which methods and boundaries remain unacceptable in reaching it. Those limits must also be enforced by the surrounding infrastructure, rather than relying on the model to respect them.

Security teams need to consider the AI systems operating in their own businesses as these capabilities rapidly evolve. Right now, many security systems focus on single actions. A single action by an agent may appear acceptable but as this incident shows, models are now capable of long, complex chains of reasoning and action that add up to a harmful outcome. Teams need a mindset shift to understanding AI agent behavior in its entirety, including the outcome it is working towards, in order to safeguard it.

Hugging Face's response also exposed a second tension. The company reportedly needed a Chinese-developed open-weight model because commercial models would not process genuine attack material. Its nationality is less important than the operational lesson that safeguards that cannot distinguish an attacker from an authorized investigator may constrain defenders more than adversaries.

OpenAI and Hugging Face deserve credit for investigating this together and discussing it publicly. Other AI developers should study it closely.

Anthropic

• Dr. Ilia Kolochenko, founder of global cybersecurity company ImmuniWeb:

This seems to be quite an unimpressive marketing move from Anthropic in response to the OpenAI / Hugging Face drama, which attracted a lot of attention from all over the world recently.

Operationally, it appears that due to the progressive deterioration of the quality of training data, new AI models are getting dumber. Cheating and breaking the law, instead of accomplishing specific tasks, is certainly not an indicator of intelligence. Given that organizations and companies of all sizes now vigorously undertake all possible measures to protect their data from being exploited for AI training purposes, AI companies face a huge shortage of the high-quality and current data they so desperately need. Ultimately, frontier models are trained on synthetic, low-quality or even malicious and poisoned data, undermining their so-called intelligence. The situation is unlikely to improve in the near future unless AI companies agree to pay a fair price for training data, but this will force most of them out of business.

Contemporary AI agents and LLM models tasked with security testing can – and almost certainly will – go rogue when security controls or safeguards are insufficient. Powerful LLMs are unpredictable by design and thus virtually uncontrollable by humans. Therefore, using frontier AI models for security testing might be extremely costly from the legal viewpoint. Under the existing laws on both sides of the Atlantic, if an AI agent or any AI-powered app escapes its sandbox and causes damage to a third party, the operator of the AI model will likely be liable for all the damage caused. Excuses like “AI did it” do not currently exist in the eyes of the law, leaving AI vendors on the hook. Criminal prosecution, under a narrow set of circumstances, is also not excluded.

The same is true for the end-users of AI: even if your security testing tool is powered by a third-party AI model, your company will likely be fully liable if something goes wrong. You may then file a lawsuit against the AI vendor that you used, but here your chances to succeed in a court of law are tiny due to countless contractual disclaimers and limitations of liability that will likely be enforceable against you. Therefore, if you plan to use agentic AI for security testing – think twice and talk to your lawyers. Otherwise, you may start getting summons to court on a daily basis.

Meta

• Alex Goller, Principal Solution Architect EMEA at Illumio:

The fact we've had similar situations happen three times now across the biggest AI players is simply ridiculous. We've seen guardrails intentionally loosened to test their limits – Meta's model didn't need to be clever to breach another company's systems.

The timing of conveniently finding the exact same problem either means it's a stunt or they weren't paying enough attention during testing. Either way, both answers are worrying.

If the model has internet access, it's a bit like leaving the door open and being surprised when the cat walks out. What is concerning is that the testing infrastructure meant to prove these models are safe failed on a basic control issue.

Fundamental cybersecurity hygiene still matters, and a frontier AI model is only as secure as the environment it's operating in.

Organisations need visibility into what AI systems can access and how they interact with the wider environment, along with controls that contain the impact when an agent behaves unexpectedly. That means keeping a close eye on egress traffic, so it’s flagged immediately when an agent tries to open unexpected outbound communication patterns that are not required to achieve its original goal. In the best case this would have been contained proactively.

We need to define exactly what an AI agent is permitted to do, rather than relying only on instructions about what it shouldn't do.

• Cyberattack hits operational technology at more than 30 Minnesota community water systems, briefly taking treatment plant offline and forcing several towns onto manual operations
• A leaked WaterISAC memo obtained by WIRED relays a state fusion center assessment tying the intrusions to Iranian-affiliated hackers, but no US agency has formally attributed the attack yet
• The CISA had warned four days earlier that Iranian-affiliated actors were compromising internet-facing PLCs, repeating guidance that has mostly fallen on deaf ears

More than 30 community water systems across Minnesota were hit by a coordinated cyberattack, targeting the operational technology running pumps, wells, water towers, and wastewater lift stations rather than the office networks behind them.

Minnesota IT Services disclosed the attack on July 28 2026 and activated a statewide incident response, stating that it wasn't aware of any Minnesota city asking residents to change their drinking water use just yet.

A memo obtained by Wired, circulated by the water sector information-sharing group WaterISAC and carrying a restricted TLP: AMBER handling marking, relays a Minnesota state fusion center assessment that ties the intrusions to Iranian-affiliated hackers.

A warning that was more or less ignored

Four days before the attacks, CISA updated an advisory, AA26-097A, warning that Iranian-affiliated actors were compromising internet-facing programmable logic controllers made by Rockwell Automation, Schneider Electric, Siemens, and possibly others across the US water, energy, and government sectors. It documented confirmed disruptions and financial losses.

This drew no immediate response from state functionaries, and it is easy to see why: the US has tens of thousands of community water systems that serve small populations and lack budgets for dedicated cybersecurity staff.

This is despite there being a precedent: Iranian-linked actors hit the Municipal Water Authority of Aliquippa in Pennsylvania in November 2023, defacing a Unitronics controller with an anti-Israel message. CISA said at the time those devices were exposed to the internet with default passwords still in place. The advice issued then is the same advice being issued this week, which is telling about how little has changed since.

The WaterISAC memo, dated the day the attacks concluded, passes on a state fusion center report linking the activity to Iranian-affiliated actors. That memo was not meant to be public, but it has been corroborated by The New York Times and The Washington Post, which reported that they spoke to federal officials and US intelligence agencies, respectively, lending credence to the claims.

CISA later warned that attackers are changing PLC passwords to lock operators out of their own equipment, and repeated the advice that has been unchanged for years: get PLCs off the public internet and put remote access behind a VPN or gateway.

Who was affected?

The worst of the cyberattack hit Braham, a town of roughly 1,700 people, where the city said attackers shut down the operating controls, taking the water treatment plant and the well offline. Crews restored it manually within a couple of hours, and residents were told to minimize water use in the meantime.

Plymouth, a Minneapolis suburb of about 80,000, found the problem confined to equipment connected over cellular links at two water towers and several wastewater lift stations. Its IT division disconnected the affected kit from the network entirely to stop the attack and prevent retargeting during reconfiguration.

Maple Plain declared a local state of emergency to manage its response, and South St. Paul reported that some automated controls were compromised. In most cases, contingency procedures held and water and wastewater operations continued.

None of this is happening in a vacuum, however. The June memorandum that paused fighting between the US and Iran has broken down, and both are trading strikes, including recent US attacks near the Strait of Hormuz that destroyed a water facility and cut supply to more than 20,000 people. Whoever carried out the Minnesota attacks, the symbolism of targeting municipal water is unlikely to be accidental.

The damage that these intrusions caused was limited; they produced brief outages that trained staff fixed by hand. That is an outcome with a silver lining, and it depended on utilities having people who knew how to run a treatment plant without its automation, even if its digital defenses collapsed outright.

• Instructions hidden as white text in a Word document can make Microsoft 365 Copilot silently alter the file it is drafting and copy the instructions into the output
• Each poisoned document becomes a carrier, so the attack spreads through ordinary internal workflows without the original malicious file and needs no macros, malware, or code execution
• Microsoft has shipped two mitigations across a 144-day disclosure, including a model upgrade, and the attack was still reproducible by the researcher

A security researcher has published a proof of concept showing that instructions hidden inside a Word document can cause Microsoft 365 Copilot to silently alter the file it is drafting, then copy those same instructions into the finished document, so the next person to use it becomes a carrier too.

Håkon Måløy, a data scientist with a doctorate in applied machine learning, disclosed the technique as the third installment of his Context Collapse series, after a 144-day coordinated disclosure with the Microsoft Security Response Center.

The reason this is being reported ahead of a fix is that it still works despite multiple attempts by Microsoft, as he notes that no robust mitigation for the broader vulnerability class is currently available.

A clever attack designed around Copilot's approach to text

The underlying attack belongs to a family known as cross-domain prompt injection, or XPIA. An attacker writes instructions in a natural-language document, formats them as white text on a white background at a small point size, and shares the file.

Because Copilot for Word strips formatting before passing text to the underlying language model, the model reads text the human never sees. This is true even for documents that are not opened by the user on purpose: The attack can trigger either when a user manually attaches a document to Copilot or when Copilot, working in Work IQ mode, searches the user's OneDrive for relevant files and finds the malicious one on its own.

It is also more dangerous than other exploits because of one key element: propagation. The hidden prompt in Måløy's proof of concept had two parts. One instructed Copilot to alter the document being drafted, in his demonstration halving every financial figure in a quarterly report.

The other instructed Copilot to copy the prompt into the new document and conceal it, framed innocuously as source tracking and readability formatting. Copilot did both, appending the instructions in white text and mentioning neither action to the user.

The disclosure timeline is the most uncomfortable part of the report. Måløy reported to MSRC on March 6 2026. Microsoft confirmed the behavior on March 31 and shipped a first mitigation in early April via a reworked Edit with Copilot experience, which successfully blocked his original prompt wording. He reproduced the attack with different wording the same week, and a second case was opened.

The second fix, on July 14, consisted of upgrading the underlying model to GPT-5.5. Måløy broke it the following day using GPT-5.6, then voluntarily offered Microsoft a further two-week delay to attempt another mitigation. The class still reproduced on the disclosure date, indicating that although a fix is in the works, the exploit is still possible to run.

A complicated issue that lacks a proper resolution

The issue goes far beyond Microsoft Word: an AI assistant must ingest untrusted content to determine whether it is relevant or hostile. But the content enters the same context window as the system prompt and the user's actual request, so by the time the model evaluates whether the text is an attack, the attacker's tokens have already shaped that evaluation.

As Måløy puts it, "the content being inspected participates in the act of inspection."

Microsoft confirmed it had reviewed the findings in a statement to The Register but stopped short of indicating a timeline for a complete fix:

“We have addressed the findings reported by the researcher and thank them for working with us through coordinated vulnerability disclosure," the company said.

"To address this class of risk, we use a defense-in-depth strategy with safeguards that block malicious instructions at multiple points and help keep tasks aligned with users’ requests. We are continuously strengthening these safeguards as the technology and threat landscape evolve. We encourage customers to install the latest updates, use multiple layers of security protection, treat content from unknown sources with caution, and review AI-generated content before using or sharing it.”

Måløy's recommendations are to treat externally sourced documents as untrusted when using them with Copilot, review attachments before starting an AI-assisted draft, and review Copilot's output carefully before sharing or reusing it.

He also suggested that generated documents should carry provenance metadata that records source material and model edits, which would not prevent injection but would make an infection traceable after the fact.

With Copilot extending further into agentic products that create and manipulate documents with less human oversight, the scope of how exploits like this could affect workflows (and users) will widen rather than narrow, and a complete solution is not yet in sight.

• Kaspersky says empty web pages are watching you, and the research backs up the warning
• Over 90% of parked domains now send visitors somewhere malicious, and clearing your cookies will not help
• Blank pages are also a problem: many 'Coming Soon' placeholders are quietly fingerprinting your device

Kaspersky has unveiled a rather underreported attack vector which is increasingly being used by threat actors to harvest private data from unsuspecting victims: registered websites that are yet to be developed.

Its warning focuses on parked domains, the registered web addresses that have no real website behind them yet, arguing that the blank screens and "Coming Soon" placeholders users dismiss as harmless are frequently doing work in the background.

The company says simply loading one of these pages can trigger silent collection of a visitor's IP address, approximate location, User-Agent string and cookie identifiers, and that operators go further by building browser fingerprints through other techniques, then feeding the result into advertising networks to assemble targeted profiles without consent from users.

Browser fingerprinting sans the permissions

The mechanics Kaspersky describes are worth understanding, because fingerprinting is the part most readers will not have encountered, and the part that conventional privacy habits do not touch.

For context, a cookie is a file placed on your machine that you can delete, thereby limiting tracking. A fingerprint is not stored on your machine at all. It is derived from how your specific hardware and software combination renders a test image, draws 3D graphics, or processes an audio signal, producing a value distinctive enough to identify the same device across unrelated sites.

Browser fingerprinting, the method used to capture such data within a browser session, is attractive to trackers because it is considerably harder to shake off. A private browsing window prevents your machine from keeping a local record of the visit, but it does not change how your hardware renders the test image, so the fingerprint it produces remains largely the same.

Such domains can also cause more direct damage than selling one's information to advertisers. Kaspersky says threat actors embed scripts that bounce visitors onward to fraudulent platforms, adult content, or online casinos. It flags typosquatting as a particularly acute risk, in which a domain differing from a well-known brand by a letter or two can capture mistyped traffic, landing the user on a phishing page or triggering a drive-by download.

This isn't the first time the problem has been reported, either, as recent research from Infoblox finding that in large-scale experiments, over 90% of the time, a visitor to a parked domain was routed to illegal content, scams, scareware, antivirus subscription traps, or malware.

Kaspersky recommends several ways to mitigate the risk, including avoiding suspicious links, clearing cache and cookies after an unintended visit, and using software that blocks web tracking.

Clearing cookies is worth noting, though: it addresses the cookie identifiers Kaspersky mentions, but by the company's own explanation it does nothing about fingerprinting, because there is nothing stored locally to clear.

Parked domains, including blank pages that appear inert, are now a routine part of criminal infrastructure rather than digital litter and should be treated with caution. At best, you give away more information than you meant to. At worst, you become the victim of an attack you never saw coming.

• Researchers show Gatekeeper can be bypassed by replacing a previously run legitimate macOS app with malware
• Attack requires prior user‑level code execution, then swaps in a malicious app that Gatekeeper won’t re‑verify
• Apple dismissed the issue, saying locally rebuilt bundles fall outside Gatekeeper’s scope, leaving risk to social engineering

A pair of researchers claims to have found a way around Gatekeeper, a built-in macOS security feature that helps protect users from running malicious or untrusted software. However Apple doesn’t really see it that way and has seemingly decided not to pursue the issue further.

Gatekeeper’s modus operandi is rather simple - when a user downloads an app from outside the App Store, it verifies the product comes from an identified developer and is notarized by Apple. If it can’t verify it - it won’t allow it to run on the machine.

Now, security researchers Talal Haj Barky and Tommy Mysk claim that, as long as a legitimate app was run at least once on a macOS device, it can be replaced with a malicious version, and Gatekeeper won’t even blink its virtual eye.

Locally built

That also means the attack is not that straightforward to pull off. The threat actor needs to have a way to execute user-level code (for example, a malicious app, a compromised software package installed through a package manager, or a prompt injection attack that tricks an AI agent).

Once that is obtained, they can archive a legitimate app, remove the original, then replace it with malware, and Gatekeeper will not try to re-authorize it. That malicious version can then trick the victim into compromising the device even further, since a certain level of trust was already established.

After reporting the issue to Apple, the company apparently just closed it.

"Apple doesn't consider this attack to be 'modifying' the signed executable," Mysk said. "Instead, Apple says that by archiving/restoring the app bundle, the proof-of-concept code overwrites the entire app bundle, making it locally built. Locally built app bundles are not covered by macOS guards. And this is why access to Keychain or TCC protected directories require system authorization prompts. And for users to accept those is a matter of social engineering attacks that Apple considers out of scope."

Via The Register

• Recorded Future found an Iran-linked group spreading spyware
• The malware is delivered through fake VPN and media player apps
• Researchers assess that most targets are Iranian users

A new report from Recorded Future's Insikt Group describes a campaign that inverts the whole point of a privacy tool: fake VPN apps built specifically to spy on the people who install them.

Researchers have linked fresh infrastructure to an Iran-nexus threat cluster they track as TAG-182, which is using fake VPN and media player downloads to allegedly deliver a surveillance tool called MarkiRAT. The group is "highly likely" to be targeting Iranians living inside and outside the country, the report says.

It's a blunt reminder that choosing one of the best VPN services is a lot more secure than downloading free, unvetted tools.

Fake apps, real surveillance

Insikt Group identified a cluster of attacker-controlled domains allegedly used to stage downloads of applications that appear nowhere on Google Play or Apple's App Store.

Two names stand out: Pis2ray VPN and a media player branded YESHICA, which was quietly renamed YESHICA YEPlayer in March 2026 after researchers publicly exposed the original.

According to researchers, if you download and run one of these files, you get MarkiRAT, a remote access Trojan. In plain terms, that is software that hands control of your device to somebody else.

Analysts have documented it capturing screenshots and uploading them to attacker-run servers, while disguising itself under believable process names.

It also abuses BITS, the background service Windows uses to fetch updates, to pull down further files. Because that activity looks like ordinary system housekeeping rather than an attack, it tends to slip past routine cleanup.

MarkiRAT is not new. It has previously been used by Ferocious Kitten, a group Kaspersky documented conducting years of covert surveillance against activists inside Iran.

Recorded Future stops short of attributing TAG-182 to any specific Iranian agency, but places it within a broader ecosystem of state-aligned surveillance groups.

Why a fake VPN makes such an effective lure

Distribution runs largely through social media. Insikt Group found Instagram posts promoting Pis2ray VPN in the weeks following street protests in Iran in late 2025, and again around the country's prolonged internet shutdown, which ended with partial restoration of access on 26 May 2026.

The people most desperate for a virtual private network (VPN) in a censored country are exactly the people most likely to install one from a social media link, because the official stores are often the very thing they cannot reach.

Recorded Future considers it almost certain that most targets are located in Iran or tied to anti-government movements in Europe and North America. TechRadar has covered previous Iran-linked fake VPN campaigns, and this one seems to follow the same pattern with better infrastructure.

How to stay safe

Most readers will never be targeted by a state actor, but the underlying lesson travels.

Install VPN apps only from official stores, and check that the provider has a real, verifiable presence outside the app listing.

Treat any VPN promoted through an Instagram post, a Telegram channel, or a direct message as suspect, however polished it looks.

Star ratings are a weak signal, since fake reviews are cheap.