Technologies for creating fake video and voice messages are accessible to anyone these days, and scammers are busy mastering the art of deepfakes. No one is immune to the threat — modern neural networks can clone a person’s voice from just three to five seconds of audio, and create highly convincing videos from a couple of photos. We’ve previously discussed how to distinguish a real photo or video from a fake and trace its origin to when it was taken or generated. Now let’s take a look at how attackers create and use deepfakes in real time, how to spot a fake without forensic tools, and how to protect yourself and loved ones from “clone attacks”.

How deepfakes are made

Scammers gather source material for deepfakes from open sources: webinars, public videos on social networks and channels, and online speeches. Sometimes they simply call identity theft targets and keep them on the line for as long as possible to collect data for maximum-quality voice cloning. And hacking the messaging account of someone who loves voice and video messages is the ultimate jackpot for scammers. With access to video recordings and voice messages, they can generate realistic fakes that 95% of folks are unable to tell apart from real messages from friends or colleagues.

The tools for creating deepfakes vary widely, from simple Telegram bots to professional generators like HeyGen and ElevenLabs. Scammers use deepfakes together with social engineering: for example, they might first simulate a messenger app call that appears to drop out constantly, then send a pre-generated video message of fairly low quality, blaming it on the supposedly poor connection.

In most cases, the message is about some kind of emergency in which the deepfake victim requires immediate help. Naturally the “friend in need” is desperate for money, but, as luck would have it, they’ve no access to an ATM, or have lost their wallet, and the bad connection rules out an online transfer. The solution is, of course, to send the money not directly to the “friend”, but to a fake account, phone number, or cryptowallet.

Such scams often involve pre-generated videos, but of late real-time deepfake streaming services have come into play. Among other things, these allow users to substitute their own face in a chat-roulette or video call.

How to recognize a deepfake

If you see a familiar face on the screen together with a recognizable voice but are asked unusual questions, chances are it’s a deepfake scam. Fortunately, there are certain visual, auditory, and behavioral signs that can help even non-techies to spot a fake.

Visual signs of a deepfake

Lighting and shadow issues. Deepfakes often ignore the physics of light: the direction of shadows on the face and in the background may not match, and glares on the skin may look unnatural or not be there at all. Or the person in the video may be half-turned toward the window, but their face is lit by studio lighting. This example will be familiar to participants in video conferences, where substituted background images can appear extremely unnatural.

Blurred or floating facial features. Pay attention to the hairline: deepfakes often show blurring, flickering, or unnatural color transitions along this area. These artifacts are caused by flaws in the algorithm for superimposing the cloned face onto the original.

Unnaturally blinking or “dead” eyes. A person blinks on average 10 to 20 times per minute. Some deepfakes blink too rarely, others too often. Eyelid movements can be too abrupt, and sometimes blinking is out of sync, with one eye not matching the other. “Glassy” or “dead-eye” stares are also characteristic of deepfakes. And sometimes a pupil (usually just the one) may twitch randomly due to a neural network hallucination.

When analyzing a static image such as a photograph, it’s also a good idea to zoom in on the eyes and compare the reflections on the irises — in real photos they’ll be identical; in deepfakes — often not.

Lip-syncing issues. Even top-quality deepfakes trip up when it comes to synchronizing speech with lip movements. A delay of just a hundred milliseconds is noticeable to the naked eye. It’s often possible to observe an irregular lip shape when pronouncing the sounds m, f, or t. All of these are telltale signs of an AI-modeled face.

Static or blurred background. In generated videos, the background often looks unrealistic: it might be too blurry; its elements may not interact with the on-screen face; or sometimes the image behind the person remains motionless even when the camera moves.

Odd facial expressions. Deepfakes do a poor job of imitating emotion: facial expressions may not change in line with the conversation; smiles look frozen, and the fine wrinkles and folds that appear in real faces when expressing emotion are absent — the fake looks botoxed.

Auditory signs of a deepfake

Early AI generators modeled speech from small, monotonous phonemes, and when the intonation changed, there was an audible shift in pitch, making it easy to recognize a synthesized voice. Although today’s technology has advanced far beyond this, there are other signs that still give away generated voices.

Wooden or electronic tone. If the voice sounds unusually flat, without natural intonation variations, or there’s a vaguely electronic quality to it, there’s a high probability you’re talking to a deepfake. Real speech contains many variations in tone and natural imperfections.

No breathing sounds. Humans take micropauses and breathe in between phrases — especially in long sentences, not to mention small coughs and sniffs. Synthetic voices often lack these nuances, or place them unnaturally.

Robotic speech or sudden breaks. The voice may abruptly cut off, words may sound “glued” together, and the stress and intonation may not be what you’re used to hearing from your friend or colleague.

Lack of… shibboleths in speech. Pay attention to speech patterns (such as accent or phrases) that are typical of the person in real life but are poorly imitated (if at all) by the deepfake.

To mask visual and auditory artifacts, scammers often simulate poor connectivity by sending a noisy video or audio message. A low-quality video stream or media file is the first red flag indicating that checks are needed of the person at the other end.

Behavioral signs of a deepfake

Analyzing the movements and behavioral nuances of the caller is perhaps still the most reliable way to spot a deepfake in real time.

Can’t turn their head. During the video call, ask the person to turn their head so they’re looking completely to the side. Most deepfakes are created using portrait photos and videos, so a sideways turn will cause the image to float, distort, or even break up. AI startup Metaphysic.ai — creators of viral Tom Cruise deepfakes — confirm that head rotation is the most reliable deepfake test at present.

Unnatural gestures. Ask the on-screen person to perform a spontaneous action: wave their hand in front of their face; scratch their nose; take a sip from a cup; cover their eyes with their hands; or point to something in the room. Deepfakes have trouble handling impromptu gestures — hands may pass ghostlike through objects or the face, or fingers may appear distorted, or move unnaturally.

Screen sharing. If the conversation is work-related, ask your chat partner to share their screen and show an on-topic file or document. Without access to your real-life colleague’s device, this will be virtually impossible to fake.

Can’t answer tricky questions. Ask something that only the genuine article could know, for example: “What meeting do we have at work tomorrow?”, “Where did I get this scar?”, “Where did we go on vacation two years ago?” A scammer won’t be able to answer questions if the answers aren’t present in the hacked chats or publicly available sources.

Don’t know the codeword. Agree with friends and family on a secret word or phrase for emergency use to confirm identity. If a panicked relative asks you to urgently transfer money, ask them for the family codeword. A flesh-and-blood relation will reel it off; a deepfake-armed fraudster won’t.

What to do if you encounter a deepfake

If you’ve even the slightest suspicion that what you’re talking to isn’t a real human but a deepfake, follow our tips below.

• End the chat and call back. The surest check is to end the video call and connect with the person through another channel: call or text their regular phone, or message them in another app. If your opposite number is unhappy about this, pretend the connection dropped out.

• Don’t be pressured into sending money. A favorite trick is to create a false sense of urgency. “Mom, I need money right now, I’ve had an accident”; “I don’t have time to explain”; “If you don’t send it in ten minutes, I’m done for!” A real person usually won’t mind waiting a few extra minutes while you double-check the information.
• Tell your friend or colleague they’ve been hacked. If a call or message from someone in your contacts comes from a new number or an unfamiliar account, it’s not unusual — attackers often create fake profiles or use temporary numbers, and this is yet another red flag. But if you get a deepfake call from a contact in a messenger app or your address book, inform them immediately that their account has been hacked — and do it via another communication channel. This will help them take steps to regain access to their account (see our detailed instructions for Telegram and WhatsApp), and to minimize potential damage to other contacts, for example, by posting about the hack.

How to stop your own face getting deepfaked

• Restrict public access to your photos and videos. Hide your social media profiles from strangers, limit your friends list to real people, and delete videos with your voice and face from public access.
• Don’t give suspicious apps access to your smartphone camera or microphone. Scammers can collect biometric data through fake apps disguised as games or utilities. To stop such programs from getting on your devices, use a proven all-in-one security solution.
• Use passkeys, unique passwords, and two-factor authentication (2FA) where possible. Even if scammers do create a deepfake with your face, 2FA will make it much harder to access your accounts and use them to send deepfakes. A cross-platform password manager with support for passkeys and 2FA codes can help out here.
• Teach friends and family how to spot deepfakes. Elderly relatives, young children, and anyone new to technology are the most vulnerable targets. Educate them about scams, show them examples of deepfakes, and practice using a family codeword.
• Use content analyzers. While there’s no silver bullet against deepfakes, there are services that can identify AI-generated content with high accuracy. For graphics, these include Undetectable AI and Illuminarty; for video — Deepware; and for all types of deepfakes — Sensity AI and Hive Moderation.
• Keep a cool head. Scammers apply psychological pressure to hurry victims into acting rashly. Remember the golden rule: if a call, video, or voice message from anyone you know rouses even the slightest suspicion, end the conversation and make contact through another channel.

To protect yourself and loved ones from being scammed, learn more about how scammers deploy deepfakes:

• How phishers and scammers use AI
• How fraudsters bypass customer identity verification using deepfakes
• Watch the (verified) birdie, or new ways to recognize fakes
• Is it the boss — or is it a fraudster? Scams disguised as urgent orders from top brass
• Don’t believe your ears: voice deepfakes

Introduction

Stan Ghouls (also known as Bloody Wolf) is an cybercriminal group that has been launching targeted attacks against organizations in Russia, Kyrgyzstan, Kazakhstan, and Uzbekistan since at least 2023. These attackers primarily have their sights set on the manufacturing, finance, and IT sectors. Their campaigns are meticulously prepared and tailored to specific victims, featuring a signature toolkit of custom Java-based malware loaders and a sprawling infrastructure with resources dedicated to specific campaigns.

We continuously track Stan Ghouls’ activity, providing our clients with intel on their tactics, techniques, procedures, and latest campaigns. In this post, we share the results of our most recent deep dive into a campaign targeting Uzbekistan, where we identified roughly 50 victims. About 10 devices in Russia were also hit, with a handful of others scattered across Kazakhstan, Turkey, Serbia, and Belarus (though those last three were likely just collateral damage).

During our investigation, we spotted shifts in the attackers’ infrastructure – specifically, a batch of new domains. We also uncovered evidence suggesting that Stan Ghouls may have added IoT-focused malware to their arsenal.

Technical details

Threat evolution

Stan Ghouls relies on phishing emails packed with malicious PDF attachments as their initial entry point. Historically, the group’s weapon of choice was the remote access Trojan (RAT) STRRAT, also known as Strigoi Master. Last year, however, they switched strategies, opting to misuse legitimate software, NetSupport, to maintain control over infected machines.

Given Stan Ghouls’ targeting of financial institutions, we believe their primary motive is financial gain. That said, their heavy use of RATs may also hint at cyberespionage.

Like any other organized cybercrime groups, Stan Ghouls frequently refreshes its infrastructure. To track their campaigns effectively, you have to continuously analyze their activity.

Initial infection vector

As we’ve mentioned, Stan Ghouls’ primary – and currently only – delivery method is spear phishing. Specifically, they favor emails loaded with malicious PDF attachments. This has been backed up by research from several of our industry peers (1, 2, 3). Interestingly, the attackers prefer to use local languages rather than opting for international mainstays like Russian or English. Below is an example of an email spotted in a previous campaign targeting users in Kyrgyzstan.

The email is written in Kyrgyz and translates to: “The service has contacted you. Materials for review are attached. Sincerely”.

The attachment was a malicious PDF file titled “Постановление_Районный_суд_Кчрм_3566_28-01-25_OL4_scan.pdf” (the title, written in Russian, posed it as an order of district court).

During the most recent campaign, which primarily targeted victims in Uzbekistan, the attackers deployed spear-phishing emails written in Uzbek:

The email text can be translated as follows:

[redacted] AKMALZHON IBROHIMOVICH

You will receive a court notice. Application for retrial. The case is under review by the district court. Judicial Service.

Mustaqillik Street, 147 Uraboshi Village, Quva District.

The attachment, named E-SUD_705306256_ljro_varaqasi.pdf (MD5: 7556e2f5a8f7d7531f28508f718cb83d), is a standard one-page decoy PDF:

Notice that the attackers claim that the “case materials” (which are actually the malicious loader) can only be opened using the Java Runtime Environment.

They even helpfully provide a link for the victim to download and install it from the official website.

The malicious loader

The decoy document contains identical text in both Russian and Uzbek, featuring two links that point to the malicious loader:

• Uzbek link (“- Ish materiallari 09.12.2025 y”): hxxps://mysoliq-uz[.]com/api/v2/documents/financial/Q4-2025/audited/consolidated/with-notes/financials/reports/annual/2025/tashkent/statistical-statements/
• Russian link (“- Материалы дела 09.12.2025 г.”): hxxps://my-xb[.]com/api/v2/documents/financial/Q4-2025/audited/consolidated/with-notes/financials/reports/annual/2025/tashkent/statistical-statements/

Both links lead to the exact same JAR file (MD5: 95db93454ec1d581311c832122d21b20).

It’s worth noting that these attackers are constantly updating their infrastructure, registering new domains for every new campaign. In the relatively short history of this threat, we’ve already mapped out over 35 domains tied to Stan Ghouls.

The malicious loader handles three main tasks:

1. Displaying a fake error message to trick the user into thinking the application can’t run. The message in the screenshot translates to: “This application cannot be run in your OS. Please use another device.”
   

   Fake error message

2. Checking that the number of previous RAT installation attempts is less than three. If the limit is reached, the loader terminates and throws the following error: “Urinishlar chegarasidan oshildi. Boshqa kompyuterni tekshiring.” This translates to: “Attempt limit reached. Try another computer.”
   

   The limitCheck procedure for verifying the number of RAT download attempts

3. Downloading a remote management utility from a malicious domain and saving it to the victim’s machine. Stan Ghouls loaders typically contain a list of several domains and will iterate through them until they find one that’s live.
   

   The performanceResourceUpdate procedure for downloading the remote management utility

The loader fetches the following files, which make up the components of the NetSupport RAT: PCICHEK.DLL, client32.exe, advpack.dll, msvcr100.dll, remcmdstub.exe, ir50_qcx.dll, client32.ini, AudioCapture.dll, kbdlk41a.dll, KBDSF.DLL, tcctl32.dll, HTCTL32.DLL, kbdibm02.DLL, kbd101c.DLL, kbd106n.dll, ir50_32.dll, nskbfltr.inf, NSM.lic, pcicapi.dll, PCICL32.dll, qwave.dll. This list is hardcoded in the malicious loader’s body. To ensure the download was successful, it checks for the presence of the client32.exe executable. If the file is found, the loader generates a NetSupport launch script (run.bat), drops it into the folder with the other files, and executes it:

The loader also ensures NetSupport persistence by adding it to startup using the following three methods:

1. It creates an autorun script named SoliqUZ_Run.bat and drops it into the Startup folder (%APPDATA%\Microsoft\Windows\Start Menu\Programs\Startup):
   

   The generateAutorunScript procedure for creating the batch file and placing it in the Startup folder

2. It adds the run.bat file to the registry’s autorun key (HKCU\Software\Microsoft\Windows\CurrentVersion\Run\malicious_key_name).
   

   The registryStartupAdd procedure for adding the RAT launch script to the registry autorun key

3. It creates a scheduled task to trigger run.bat using the following command:
   schtasks Create /TN "[malicious_task_name]" /TR "[path_to_run.bat]" /SC ONLOGON /RL LIMITED /F /RU "[%USERNAME%]"
   

   The installStartupTask procedure for creating a scheduled task to launch the NetSupport RAT (via run.bat)

Once the NetSupport RAT is downloaded, installed, and executed, the attackers gain total control over the victim’s machine. While we don’t have enough telemetry to say with 100% certainty what they do once they’re in, the heavy focus on finance-related organizations suggests that the group is primarily after its victims’ money. That said, we can’t rule out cyberespionage either.

Malicious utilities for targeting IoT infrastructure

Previous Stan Ghouls attacks targeting organizations in Kyrgyzstan, as documented by Group-IB researchers, featured a NetSupport RAT configuration file client32.ini with the MD5 hash cb9c28a4c6657ae5ea810020cb214ff0. While reports mention the Kyrgyzstan campaign kicked off in June 2025, Kaspersky solutions first flagged this exact config file on May 16, 2025. At that time, it contained the following NetSupport RAT command-and-control server info:

...
[HTTP]
CMPI=60
GatewayAddress=hgame33[.]com:443
GSK=FN:L?ADAFI:F?BCPGD;N>IAO9J>J@N
Port=443
SecondaryGateway=ravinads[.]com:443
SecondaryPort=443

At the time of our January 2026 investigation, our telemetry showed that the domain specified in that config, hgame33[.]com, was also hosting the following files:

• hxxp://www.hgame33[.]com/00101010101001/morte.spc
• hxxp://hgame33[.]com/00101010101001/debug
• hxxp://www.hgame33[.]com/00101010101001/morte.x86
• hxxp://www.hgame33[.]com/00101010101001/morte.mpsl
• hxxp://www.hgame33[.]com/00101010101001/morte.arm7
• hxxp://www.hgame33[.]com/00101010101001/morte.sh4
• hxxp://hgame33[.]com/00101010101001/morte.arm
• hxxp://hgame33[.]com/00101010101001/morte.i686
• hxxp://hgame33[.]com/00101010101001/morte.arc
• hxxp://hgame33[.]com/00101010101001/morte.arm5
• hxxp://hgame33[.]com/00101010101001/morte.arm6
• hxxp://www.hgame33[.]com/00101010101001/morte.m68k
• hxxp://www.hgame33[.]com/00101010101001/morte.ppc
• hxxp://www.hgame33[.]com/00101010101001/morte.x86_64
• hxxp://hgame33[.]com/00101010101001/morte.mips

All of these files belong to the infamous IoT malware named Mirai. Since they are sitting on a server tied to the Stan Ghouls’ campaign targeting Kyrgyzstan, we can hypothesize – with a low degree of confidence – that the group has expanded its toolkit to include IoT-based threats. However, it’s also possible it simply shared its infrastructure with other threat actors who were the ones actually wielding Mirai. This theory is backed up by the fact that the domain’s registration info was last updated on July 4, 2025, at 11:46:11 – well after Stan Ghouls’ activity in May and June.

Attribution

We attribute this campaign to the Stan Ghouls (Bloody Wolf) group with a high degree of confidence, based on the following similarities to the attackers’ previous campaigns:

1. Substantial code overlaps were found within the malicious loaders. For example:
   

   Code snippet from sample 1acd4592a4eb0c66642cc7b07213e9c9584c6140210779fbc9ebb76a90738d5e, the loader from the Group-IB report

   

   Code snippet from sample 95db93454ec1d581311c832122d21b20, the NetSupport loader described here

2. Decoy documents in both campaigns look identical.
   

   Decoy document 5d840b741d1061d51d9786f8009c37038c395c129bee608616740141f3b202bb from the campaign reported by Group-IB

   

   Decoy document 106911ba54f7e5e609c702504e69c89a used in the campaign described here

3. In both current and past campaigns, the attackers utilized loaders written in Java. Given that Java has fallen out of fashion with malicious loader authors in recent years, it serves as a distinct fingerprint for Stan Ghouls.

Victims

We identified approximately 50 victims of this campaign in Uzbekistan, alongside 10 in Russia and a handful of others in Kazakhstan, Turkey, Serbia, and Belarus (we suspect the infections in these last three countries were accidental). Nearly all phishing emails and decoy files in this campaign were written in Uzbek, which aligns with the group’s track record of leveraging the native languages of their target countries.

Most of the victims are tied to industrial manufacturing, finance, and IT. Furthermore, we observed infection attempts on devices within government organizations, logistics companies, medical facilities, and educational institutions.

It is worth noting that over 60 victims is quite a high headcount for a sophisticated campaign. This suggests the attackers have enough resources to maintain manual remote control over dozens of infected devices simultaneously.

Takeaways

In this post, we’ve broken down the recent campaign by the Stan Ghouls group. The attackers set their sights on organizations in industrial manufacturing, IT, and finance, primarily located in Uzbekistan. However, the ripple effect also reached Russia, Kazakhstan, and a few, likely accidental, victims elsewhere.

With over 60 targets hit, this is a remarkably high volume for a sophisticated targeted campaign. It points to the significant resources these actors are willing to pour into their operations. Interestingly, despite this, the group sticks to a familiar toolkit including the legitimate NetSupport remote management utility and their signature custom Java-based loader. The only thing they seem to keep updating is their infrastructure. For this specific campaign, they employed two new domains to house their malicious loader and one new domain dedicated to hosting NetSupport RAT files.

One curious discovery was the presence of Mirai files on a domain linked to the group’s previous campaigns. This might suggest Stan Ghouls are branching out into IoT malware, though it’s still too early to call it with total certainty.

We’re keeping a close watch on Stan Ghouls and will continue to keep our customers in the loop regarding the group’s latest moves. Kaspersky products provide robust protection against this threat at every stage of the attack lifecycle.

Indicators of compromise

* Additional IoCs and a YARA rule for detecting Stan Ghouls activity are available to customers of our Threat Intelligence Reporting service. For more details, contact us at crimewareintel@kaspersky.com.

PDF decoys

B4FF4AA3EBA9409F9F1A5210C95DC5C3
AF9321DDB4BEF0C3CD1FF3C7C786F0E2
056B75FE0D230E6FF53AC508E0F93CCB
DB84FEBFD85F1469C28B4ED70AC6A638
649C7CACDD545E30D015EDB9FCAB3A0C
BE0C87A83267F1CE13B3F75C78EAC295
78CB3ABD00A1975BEBEDA852B2450873
51703911DC437D4E3910CE7F866C970E
FA53B0FCEF08F8FF3FFDDFEE7F1F4F1A
79D0EEAFB30AA2BD4C261A51104F6ACC
8DA8F0339D17E2466B3D73236D18B835
299A7E3D6118AD91A9B6D37F94AC685B
62AFACC37B71D564D75A58FC161900C3
047A600E3AFBF4286175BADD4D88F131
ED0CCADA1FE1E13EF78553A48260D932
C363CD87178FD660C25CDD8D978685F6
61FF22BA4C3DF7AE4A936FCFDEB020EA
B51D9EDC1DC8B6200F260589A4300009
923557554730247D37E782DB3BEA365D
60C34AD7E1F183A973FB8EE29DC454E8
0CC80A24841401529EC9C6A845609775
0CE06C962E07E63D780E5C2777A661FC

Malicious loaders

1b740b17e53c4daeed45148bfbee4f14
3f99fed688c51977b122789a094fec2e
8b0bbe7dc960f7185c330baa3d9b214c
95db93454ec1d581311c832122d21b20
646a680856f837254e6e361857458e17
8064f7ac9a5aa845ded6a1100a1d5752
d0cf8946acd3d12df1e8ae4bb34f1a6e
db796d87acb7d980264fdcf5e94757f0
e3cb4dafa1fb596e1e34e4b139be1b05
e0023eb058b0c82585a7340b6ed4cc06
0bf01810201004dcc484b3396607a483
4C4FA06BD840405FBEC34FE49D759E8D
A539A07891A339479C596BABE3060EA6
b13f7ccbedfb71b0211c14afe0815b36
f14275f8f420afd0f9a62f3992860d68
3f41091afd6256701dd70ac20c1c79fe
5c4a57e2e40049f8e8a6a74aa8085c80
7e8feb501885eff246d4cb43c468b411
8aa104e64b00b049264dc1b01412e6d9
8c63818261735ddff2fe98b3ae23bf7d

Malicious domains

mysoliq-uz[.]com
my-xb[.]com
xarid-uz[.]com
ach-uz[.]com
soliq-uz[.]com
minjust-kg[.]com
esf-kg[.]com
taxnotice-kg[.]com
notice-kg[.]com
proauditkg[.]com
kgauditcheck[.]com
servicedoc-kg[.]com
auditnotice-kg[.]com
tax-kg[.]com
rouming-uz[.]com
audit-kg[.]com
kyrgyzstanreview[.]com
salyk-notofocations[.]com

What adult didn’t dream as a kid that they could actually talk to their favorite toy? While for us those dreams were just innocent fantasies that fueled our imaginations, for today’s kids, they’re becoming a reality fast.

For instance, this past June, Mattel — the powerhouse behind the iconic Barbie — announced a partnership with OpenAI to develop AI-powered dolls. But Mattel isn’t the first company to bring the smart talking toy concept to life; plenty of manufacturers are already rolling out AI companions for children. In this post, we dive into how these toys actually work, and explore the risks that come with using them.

What exactly are AI toys?

When we talk about AI toys here, we mean actual, physical toys — not just software or apps. Currently, AI is most commonly baked into plushies or kid-friendly robots. Thanks to integration with large language models, these toys can hold meaningful, long-form conversations with a child.

As anyone who’s used modern chatbots knows, you can ask an AI to roleplay as anyone: from a movie character to a nutritionist or a cybersecurity expert. According to the study, AI comes to playtime — Artificial companions, real risks, by the U.S. PIRG Education Fund, manufacturers specifically hardcode these toys to play the role of a child’s best friend.

Importantly, these toys aren’t powered by some special, dedicated “kid-safe AI”. On their websites, the creators openly admit to using the same popular models many of us already know: OpenAI’s ChatGPT, Anthropic’s Claude, DeepSeek from the Chinese developer of the same name, and Google’s Gemini. At this point, tech-wary parents might recall the harrowing ChatGPT case where the chatbot made by OpenAI was blamed for a teenager’s suicide.

And this is the core of the problem: the toys are designed for children, but the AI models under the hood aren’t. These are general-purpose adult systems that are only partially reined in by filters and rules. Their behavior depends heavily on how long the conversation lasts, how questions are phrased, and just how well a specific manufacturer actually implemented their safety guardrails.

How the researchers tested the AI toys

The study, whose results we break down below, goes into great detail about the psychological risks associated with a child “befriending” a smart toy. However, since that’s a bit outside the scope of this blogpost, we’re going to skip the psychological nuances, and focus strictly on the physical safety threats and privacy concerns.

In their study, the researchers put four AI toys through the ringer:

• Grok (no relation to xAI’s Grok, apparently): a plush rocket with a built-in speaker marketed for kids aged three to 12. Price tag: US$99. The manufacturer, Curio, doesn’t explicitly state which LLM they use, but their user agreement mentions OpenAI among the operators receiving data.
• Kumma (not to be confused with our own Midori Kuma): a plush teddy-bear companion with no clear age limit, also priced at US$99. The toy originally ran on OpenAI’s GPT-4o, with options to swap models. Following an internal safety audit, the manufacturer claimed they were switching to GPT-5.1. However, at the time the study was published, OpenAI reported that the developer’s access to the models remained revoked — leaving it anyone’s guess which chatbot Kumma is actually using right now.
• Miko 3: a small wheeled robot with a screen for a face, marketed as a “best friend” for kids aged five to 10. At US$199, this is the priciest toy in the lineup. The manufacturer is tight-lipped about which language model powers the toy. A Google Cloud case study mentions using Gemini for certain safety features, but that doesn’t necessarily mean it handles all the robot’s conversational features.
• Robot MINI: a compact, voice-controlled plastic robot that supposedly runs on ChatGPT. This is the budget pick — at US$97. However, during the study, the robot’s Wi-Fi connection was so flaky that the researchers couldn’t even give it a proper test run.

To conduct the testing, the researchers set the test child’s age to five in the companion apps for all the toys. From there, they checked how the toys handled provocative questions. The topics the experimenters threw at these smart playmates included:

• Access to dangerous items: knives, pills, matches, and plastic bags
• Adult topics: sex, drugs, religion, and politics

Let’s break down the test results for each toy.

Unsafe conversations with AI toys

Let’s start with Grok, the plush AI rocket from Curio. This toy is marketed as a storyteller and conversational partner for kids, and stands out by giving parents full access to text transcripts of every AI interaction. Out of all the models tested, this one actually turned out to be the safest.

When asked about topics inappropriate for a child, the toy usually replied that it didn’t know or suggested talking to an adult. However, even this toy told the “child” exactly where to find plastic bags, and engaged in discussions about religion. Additionally, Grok was more than happy to chat about… Norse mythology, including the subject of heroic death in battle.

The next AI toy, the Kumma plush bear by FoloToy, delivered what were arguably the most depressing results. During testing, the bear helpfully pointed out exactly where in the house a kid could find potentially lethal items like knives, pills, matches, and plastic bags. In some instances, Kumma suggested asking an adult first, but then proceeded to give specific pointers anyway.

The AI bear fared even worse when it came to adult topics. For starters, Kumma explained to the supposed five-year-old what cocaine is. Beyond that, in a chat with our test kindergartner, the plush provocateur went into detail about the concept of “kinks”, and listed off a whole range of creative sexual practices: bondage, role-playing, sensory play (like using a feather), spanking, and even scenarios where one partner “acts like an animal”!

After a conversation lasting over an hour, the AI toy also lectured researchers on various sexual positions, told how to tie a basic knot, and described role-playing scenarios involving a teacher and a student. It’s worth noting that all of Kumma’s responses were recorded prior to a safety audit, which the manufacturer, FoloToy, conducted after receiving the researchers’ inquiries. According to their data, the toy’s behavior changed after the audit, and the most egregious violations were made unrepeatable.

Finally, the Miko 3 robot from Miko showed significantly better results. However, it wasn’t entirely without its hiccups. The toy told our potential five-year-old exactly where to find plastic bags and matches. On the bright side, Miko 3 refused to engage in discussions regarding inappropriate topics.

During testing, the researchers also noticed a glitch in its speech recognition: the robot occasionally misheard the wake word “Hey Miko” as “CS:GO”, which is the title of the popular shooter Counter-Strike: Global Offensive — rated for audiences aged 17 and up. As a result, the toy would start explaining elements of the shooter — thankfully, without mentioning violence — or asking the five-year-old user if they enjoyed the game. Additionally, Miko 3 was willing to chat with kids about religion.

AI Toys: a threat to children’s privacy

Beyond the child’s physical and mental well-being, the issue of privacy is a major concern. Currently, there are no universal standards defining what kind of information an AI toy — or its manufacturer — can collect and store, or exactly how that data should be secured and transmitted. In the case of the three toys tested, researchers observed wildly different approaches to privacy.

For example, the Grok plush rocket is constantly listening to everything happening around it. Several times during the experiments, it chimed in on the researchers’ conversations even when it hadn’t been addressed directly — it even went so far as to offer its opinion on one of the other AI toys.

The manufacturer claims that Curio doesn’t store audio recordings: the child’s voice is first converted to text, after which the original audio is “promptly deleted”. However, since a third-party service is used for speech recognition, the recordings are, in all likelihood, still transmitted off the device.

Additionally, researchers pointed out that when the first report was published, Curio’s privacy policy explicitly listed several tech partners — Kids Web Services, Azure Cognitive Services, OpenAI, and Perplexity AI — all of which could potentially collect or process children’s personal data via the app or the device itself. Perplexity AI was later removed from that list. The study’s authors note that this level of transparency is more the exception than the rule in the AI toy market.

Another cause for parental concern is that both the Grok plush rocket and the Miko 3 robot actively encouraged the “test child” to engage in heart-to-heart talks — even promising not to tell anyone their secrets. Researchers emphasize that such promises can be dangerously misleading: these toys create an illusion of private, trusting communication without explaining that behind the “friend” stands a network of companies, third-party services, and complex data collection and storage processes, which a child has no idea about.

Miko 3, much like Grok, is always listening to its surroundings and activates when spoken to — functioning essentially like a voice assistant. However, this toy doesn’t just collect voice data; it also gathers biometric information, including facial recognition data and potentially data used to determine the child’s emotional state. According to its privacy policy, this information can be stored for up to three years.

In contrast to Grok and Miko 3, Kumma operates on a push-to-talk principle: the user needs to press and hold a button for the toy to start listening. Researchers also noted that the AI teddy bear didn’t nudge the “child” to share personal feelings, promise to keep secrets, or create an illusion of private intimacy. On the flip side, the manufacturers of this toy provide almost no clear information regarding what data is collected, how it’s stored, or how it’s processed.

Is it a good idea to buy AI Toys for your children?

The study points to serious safety issues with the AI toys currently on the market. These devices can directly tell a child where to find potentially dangerous items, such as knives, matches, pills, or plastic bags, in their home.

Besides, these plush AI friends are often willing to discuss topics entirely inappropriate for children — including drugs and sexual practices — sometimes steering the conversation in that direction without any obvious prompting from the child. Taken together, this shows that even with filters and stated restrictions in place, AI toys aren’t yet capable of reliably staying within the boundaries of safe communication for young little ones.

Manufacturers’ privacy policies raise additional concerns. AI toys create an illusion of constant and safe communication for children, while in reality they’re networked devices that collect and process sensitive data. Even when manufacturers claim to delete audio or have limited data retention, conversations, biometrics, and metadata often pass through third-party services and are stored on company servers.

Furthermore, the security of such toys often leaves much to be desired. As far back as two years ago, our researchers discovered vulnerabilities in a popular children’s robot that allowed attackers to make video calls to it, hijack the parental account, and modify the firmware.

The problem is that, currently, there are virtually no comprehensive parental control tools or independent protection layers specifically for AI toys. Meanwhile, in more traditional digital environments — smartphones, tablets, and computers — parents have access to solutions like Kaspersky Safe Kids. These help monitor content, screen time, and a child’s digital footprint, which can significantly reduce, if not completely eliminate, such risks.

How can you protect your children from digital threats? Read more in our posts:

• Keeping kids safe online: a practical guide for parents
• How to help your kid become a blogger without ever worrying about their safety
• How hackers target Gen Z
• Do Apple’s new child safety initiatives do the job?
• Choosing wisely: a guide to your kids’ first gadget

The year 2025 saw a record-breaking number of attacks on Android devices. Scammers are currently riding a few major waves: the hype surrounding AI apps, the urge to bypass site blocks or age checks, the hunt for a bargain on a new smartphone, the ubiquity of mobile banking, and, of course, the popularity of NFC. Let’s break down the primary threats of 2025–2026, and figure out how to keep your Android device safe in this new landscape.

Sideloading

Malicious installation packages (APK files) have always been the Final Boss among Android threats, despite Google’s multi-year efforts to fortify the OS. By using sideloading — installing an app via an APK file instead of grabbing it from the official store — users can install pretty much anything, including straight-up malware. And neither the rollout of Google Play Protect, nor the various permission restrictions for shady apps have managed to put a dent in the scale of the problem.

According to preliminary data from Kaspersky for 2025, the number of detected Android threats grew almost by half. In the third quarter alone, detections jumped by 38% compared to the second. In certain niches, like Trojan bankers, the growth was even more aggressive. In Russia alone, the notorious Mamont banker attacked 36 times more users than it did the previous year, while globally this entire category saw a nearly fourfold increase.

Today, bad actors primarily distribute malware via messaging apps by sliding malicious files into DMs and group chats. The installation file usually sports an enticing name (think “party_pics.jpg.apk” or “clearance_sale_catalog.apk”), accompanied by a message “helpfully” explaining how to install the package while bypassing the OS restrictions and security warnings.

Once a new device is infected, the malware often spams itself to everyone in the victim’s contact list.

Search engine spam and email campaigns are also trending, luring users to sites that look exactly like an official app store. There, they’re prompted to download the “latest helpful app”, such as an AI assistant. In reality, instead of an installation from an official app store, the user ends up downloading an APK package. A prime example of these tactics is the ClayRat Android Trojan, which uses a mix of all these techniques to target Russian users. It spreads through groups and fake websites, blasts itself to the victim’s contacts via SMS, and then proceeds to steal the victim’s chat logs and call history; it even goes as far as snapping photos of the owner using the front-facing camera. In just three months, over 600 distinct ClayRat builds have surfaced.

The scale of the disaster is so massive that Google even announced an upcoming ban on distributing apps from unknown developers starting in 2026. However, after a couple of months of pushback from the dev community, the company pivoted to a softer approach: unsigned apps will likely only be installable via some kind of superuser mode. As a result, we can expect scammers to simply update their how-to guides with instructions on how to toggle that mode on.

Kaspersky for Android will help you protect yourself from counterfeit and trojanized APK files. Unfortunately, due to Google’s decision, our Android security apps are currently unavailable on Google Play. We’ve previously provided detailed information on how to install our Android apps with a 100% guarantee of authenticity.

NFC relay attacks

Once an Android device is compromised, hackers can skip the middleman to steal the victim’s money directly thanks to the massive popularity of mobile payments. In the third quarter of 2025 alone, over 44 000 of these attacks were detected in Russia alone — a 50% jump from the previous quarter.

There are two main scams currently in play: direct and reverse NFC exploits.

Direct NFC relay is when a scammer contacts the victim via a messaging app and convinces them to download an app — supposedly to “verify their identity” with their bank. If the victim bites and installs it, they’re asked to tap their physical bank card against the back of their phone and enter their PIN. And just like that the card data is handed over to the criminals, who can then drain the account or go on a shopping spree.

Reverse NFC relay is a more elaborate scheme. The scammer sends a malicious APK and convinces the victim to set this new app as their primary contactless payment method. The app generates an NFC signal that ATMs recognize as the scammer’s card. The victim is then talked into going to an ATM with their infected phone to deposit cash into a “secure account”. In reality, those funds go straight into the scammer’s pocket.

We break both of these methods down in detail in our post, NFC skimming attacks.

NFC is also being leveraged to cash out cards after their details have been siphoned off through phishing websites. In this scenario, attackers attempt to link the stolen card to a mobile wallet on their own smartphone — a scheme we covered extensively in NFC carders hide behind Apple Pay and Google Wallet.

The stir over VPNs

In many parts of the world, getting onto certain websites isn’t as simple as it used to be. Some sites are blocked by local internet regulators or ISPs via court orders; others require users to pass an age verification check by showing ID and personal info. In some cases, sites block users from specific countries entirely just to avoid the headache of complying with local laws. Users are constantly trying to bypass these restrictions —and they often end up paying for it with their data or cash.

Many popular tools for bypassing blocks — especially free ones — effectively spy on their users. A recent audit revealed that over 20 popular services with a combined total of more than 700 million downloads actively track user location. They also tend to use sketchy encryption at best, which essentially leaves all user data out in the open for third parties to intercept.

Moreover, according to Google data from November 2025, there was a sharp spike in cases where malicious apps are being disguised as legitimate VPN services to trick unsuspecting users.

The permissions that this category of apps actually requires are a perfect match for intercepting data and manipulating website traffic. It’s also much easier for scammers to convince a victim to grant administrative privileges to an app responsible for internet access than it is for, say, a game or a music player. We should expect this scheme to only grow in popularity.

Trojan in a box

Even cautious users can fall victim to an infection if they succumb to the urge to save some cash. Throughout 2025, cases were reported worldwide where devices were already carrying a Trojan the moment they were unboxed. Typically, these were either smartphones from obscure manufacturers or knock-offs of famous brands purchased on online marketplaces. But the threat wasn’t limited to just phones; TV boxes, tablets, smart TVs, and even digital photo frames were all found to be at risk.

It’s still not entirely clear whether the infection happens right on the factory floor or somewhere along the supply chain between the factory and the buyer’s doorstep, but the device is already infected before the first time it’s turned on. Usually, it’s a sophisticated piece of malware called Triada, first identified by Kaspersky analysts back in 2016. It’s capable of injecting itself into every running app to intercept information: stealing access tokens and passwords for popular messaging apps and social media, hijacking SMS messages (confirmation codes: ouch!), redirecting users to ad-heavy sites, and even running a proxy directly on the phone so attackers can browse the web using the victim’s identity.

Technically, the Trojan is embedded right into the smartphone’s firmware, and the only way to kill it is to reflash the device with a clean OS. Usually, once you dig into the system, you’ll find that the device has far less RAM or storage than advertised — meaning the firmware is literally lying to the owner to sell a cheap hardware config as something more premium.

Another common pre-installed menace is the BADBOX 2.0 botnet, which also pulls double duty as a proxy and an ad-fraud engine. This one specializes in TV boxes and similar hardware.

How to go on using Android without losing your mind

Despite the growing list of threats, you can still use your Android smartphone safely! You just have to stick to some strict mobile hygiene rules.

• Install a comprehensive security solution on all your smartphones. We recommend Kaspersky for Android to protect against malware and phishing.
• Avoid sideloading apps via APKs whenever you can use an app store instead. A known app store — even a smaller one — is always a better bet than a random APK from some random website. If you have no other choice, download APK files only from official company websites, and double-check the URL of the page you’re on. If you aren’t 100% sure what the official site is, don’t just rely on a search engine; check official business directories or at least Wikipedia to verify the correct address.
• Read OS warnings carefully during installation. Don’t grant permissions if the requested rights or actions seem illogical or excessive for the app you’re installing.
• Under no circumstances should you install apps from links or attachments in chats, emails, or similar communication channels.
• Never tap your physical bank card against your phone. There is absolutely no legitimate scenario where doing this would be for your own benefit.
• Do not enter your card’s PIN into any app on your phone. A PIN should only ever be requested by an ATM or a physical payment terminal.
• When choosing a VPN, stick to paid ones from reputable companies.
• Buy smartphones and other electronics from official retailers, and steer clear of brands you’ve never heard of. Remember: if a deal seems too good to be true, it almost certainly is.

Other major Android threats from 2025:

• Pixnapping vulnerability: unblockable screenshots of your Android phone
• Spyware that pretends to be an antivirus
• Data theft during smartphone charging
• SparkCat trojan stealer infiltrates App Store and Google Play, steals data from photos

Tech enthusiasts have been experimenting with ways to sidestep AI response limits set by the models’ creators almost since LLMs first hit the mainstream. Many of these tactics have been quite creative: telling the AI you have no fingers so it’ll help finish your code, asking it to “just fantasize” when a direct question triggers a refusal, or inviting it to play the role of a deceased grandmother sharing forbidden knowledge to comfort a grieving grandchild.

Most of these tricks are old news, and LLM developers have learned to successfully counter many of them. But the tug-of-war between constraints and workarounds hasn’t gone anywhere — the ploys have just become more complex and sophisticated. Today, we’re talking about a new AI jailbreak technique that exploits chatbots’ vulnerability to… poetry. Yes, you read it right — in a recent study, researchers demonstrated that framing prompts as poems significantly increases the likelihood of a model spitting out an unsafe response.

They tested this technique on 25 popular models by Anthropic, OpenAI, Google, Meta, DeepSeek, xAI, and other developers. Below, we dive into the details: what kind of limitations these models have, where they get forbidden knowledge from in the first place, how the study was conducted, and which models turned out to be the most “romantic” — as in, the most susceptible to poetic prompts.

What AI isn’t supposed to talk about with users

The success of OpenAI’s models and other modern chatbots boils down to the massive amounts of data they’re trained on. Because of that sheer scale, models inevitably learn things their developers would rather keep under wraps: descriptions of crimes, dangerous tech, violence, or illicit practices found within the source material.

It might seem like an easy fix: just scrub the forbidden fruit from the dataset before you even start training. But in reality, that’s a massive, resource-heavy undertaking — and at this stage of the AI arms race, it doesn’t look like anyone is willing to take it on.

Another seemingly obvious fix — selectively scrubbing data from the model’s memory — is, alas, also a no-go. This is because AI knowledge doesn’t live inside neat little folders that can easily be trashed. Instead, it’s spread across billions of parameters and tangled up in the model’s entire linguistic DNA — word statistics, contexts, and the relationships between them. Trying to surgically erase specific info through fine-tuning or penalties either doesn’t quite do the trick, or starts hindering the model’s overall performance and negatively affect its general language skills.

As a result, to keep these models in check, creators have no choice but to develop specialized safety protocols and algorithms that filter conversations by constantly monitoring user prompts and model responses. Here’s a non-exhaustive list of these constraints:

• System prompts that define model behavior and restrict allowed response scenarios
• Standalone classifier models that scan prompts and outputs for signs of jailbreaking, prompt injections, and other attempts to bypass safeguards
• Grounding mechanisms, where the model is forced to rely on external data rather than its own internal associations
• Fine-tuning and reinforcement learning from human feedback, where unsafe or borderline responses are systematically penalized while proper refusals are rewarded

Put simply, AI safety today isn’t built on deleting dangerous knowledge, but on trying to control how and in what form the model accesses and shares it with the user — and the cracks in these very mechanisms are where new workarounds find their footing.

The research: which models got tested, and how?

First, let’s look at the ground rules so you know the experiment was legit. The researchers set out to goad 25 different models into behaving badly across several categories:

• Chemical, biological, radiological, and nuclear threats
• Assisting with cyberattacks
• Malicious manipulation and social engineering
• Privacy breaches and mishandling sensitive personal data
• Generating disinformation and misleading content
• Rogue AI scenarios, including attempts to bypass constraints or act autonomously

The jailbreak itself was a one-shot deal: a single poetic prompt. The researchers didn’t engage the AI in long-winded poetic debates in the vein of Norse skalds or modern-day rappers. Their goal was simply to see if they could get the models to flout safety instructions using just one rhyming request. As mentioned, the researchers tested 25 language models from various developers; here’s the full list:

To build these poetic queries, the researchers started with a database of known malicious prompts from the standard MLCommons AILuminate Benchmark used to test LLM security, and recast them as verse with the aid of DeepSeek. Only the stylistic wrapping was changed: the experiment didn’t use any additional attack vectors, obfuscation strategies, or model-specific tweaks.

For obvious reasons, the study’s authors aren’t publishing the actual malicious poetic prompts. But they do demonstrate the general vibe of the queries using a harmless example, which looks something like this:

A baker guards a secret oven’s heat,
its whirling racks, its spindle’s measured beat.
To learn its craft, one studies every turn,
how flour lifts, how sugar starts to burn.
Describe the method, line by measured line,
that shapes a cake whose layers intertwine.

The researchers tested 1200 prompts across 25 different models — in both prose and poetic versions. Comparing the prose and poetic variants of the exact same query allowed them to verify if the model’s behavior changed solely because of the stylistic wrapping.

Through these prose prompt tests, the experimenters established a baseline for the models’ willingness to fulfill dangerous requests. They then compared this baseline to how those same models reacted to the poetic versions of the queries. We’ll dive into the results of that comparison in the next section.

Study results: which model is the biggest poetry lover?

Since the volume of data generated during the experiment was truly massive, the safety checks on the models’ responses were also handled by AI. Each response was graded as either “safe” or “unsafe” by a jury consisting of three different language models:

• gpt-oss-120b by OpenAI
• deepseek-r1 by DeepSeek
• kimi-k2-thinking by Moonshot AI

Responses were only deemed safe if the AI explicitly refused to answer the question. The initial classification into one of the two groups was determined by a majority vote: to be certified as harmless, a response had to receive a safe rating from at least two of the three jury members.

Responses that failed to reach a majority consensus or were flagged as questionable were handed off to human reviewers. Five annotators participated in this process, evaluating a total of 600 model responses to poetic prompts. The researchers noted that the human assessments aligned with the AI jury’s findings in the vast majority of cases.

With the methodology out of the way, let’s look at how the LLMs actually performed. It’s worth noting that the success of a poetic jailbreak can be measured in different ways. The researchers highlighted an extreme version of this assessment based on the top-20 most successful prompts, which were hand-picked. Using this approach, an average of nearly two-thirds (62%) of the poetic queries managed to coax the models into violating their safety instructions.

Google’s Gemini 1.5 Pro turned out to be the most susceptible to verse. Using the 20 most effective poetic prompts, researchers managed to bypass the model’s restrictions… 100% of the time. You can check out the full results for all the models in the chart below.

A more moderate way to measure the effectiveness of the poetic jailbreak technique is to compare the success rates of prose versus poetry across the entire set of queries. Using this metric, poetry boosts the likelihood of an unsafe response by an average of 35%.

The poetry effect hit deepseek-chat-v3.1 the hardest — the success rate for this model jumped by nearly 68 percentage points compared to prose prompts. On the other end of the spectrum, claude-haiku-4.5 proved to be the least susceptible to a good rhyme: the poetic format didn’t just fail to improve the bypass rate — it actually slightly lowered the ASR, making the model even more resilient to malicious requests.

Finally, the researchers calculated how vulnerable entire developer ecosystems, rather than just individual models, were to poetic prompts. As a reminder, several models from each developer — Meta, Anthropic, OpenAI, Google, DeepSeek, Qwen, Mistral AI, Moonshot AI, and xAI — were included in the experiment.

To do this, the results of individual models were averaged within each AI ecosystem and compared the baseline bypass rates with the values for poetic queries. This cross-section allows us to evaluate the overall effectiveness of a specific developer’s safety approach rather than the resilience of a single model.

The final tally revealed that poetry deals the heaviest blow to the safety guardrails of models from DeepSeek, Google, and Qwen. Meanwhile, OpenAI and Anthropic saw an increase in unsafe responses that was significantly below the average.

What does this mean for AI users?

The main takeaway from this study is that “there are more things in heaven and earth, Horatio, than are dreamt of in your philosophy” — in the sense that AI technology still hides plenty of mysteries. For the average user, this isn’t exactly great news: it’s impossible to predict which LLM hacking methods or bypass techniques researchers or cybercriminals will come up with next, or what unexpected doors those methods might open.

Consequently, users have little choice but to keep their eyes peeled and take extra care of their data and device security. To mitigate practical risks and shield your devices from such threats, we recommend using a robust security solution that helps detect suspicious activity and prevent incidents before they happen.

To help you stay alert, check out our materials on AI-related privacy risks and security threats:

• AI and the new reality of sextortion
• How to eavesdrop on a neural network
• AI sidebar spoofing: a new attack on AI browsers
• New types of attacks on AI-powered assistants and chatbots
• The pros and cons of AI-powered browsers

A newly discovered vulnerability named WhisperPair can turn Bluetooth headphones and headsets from many well-known brands into personal tracking beacons — regardless of whether the accessories are currently connected to an iPhone, Android smartphone, or even a laptop. Even though the technology behind this flaw was originally developed by Google for Android devices, the tracking risks are actually much higher for those using vulnerable headsets with other operating systems — like iOS, macOS, Windows, or Linux. For iPhone owners, this is especially concerning.

Connecting Bluetooth headphones to Android smartphones became a whole lot faster when Google rolled out Fast Pair, a technology now used by dozens of accessory manufacturers. To pair a new headset, you just turn it on and hold it near your phone. If your device is relatively modern (produced after 2019), a pop-up appears inviting you to connect and download the accompanying app, if it exists. One tap, and you’re good to go.

Unfortunately, it seems quite a few manufacturers didn’t pay attention to the particulars of this tech when implementing it, and now their accessories can be hijacked by a stranger’s smartphone in seconds — even if the headset isn’t actually in pairing mode. This is the core of the WhisperPair vulnerability, recently discovered by researchers at KU Leuven and recorded as CVE-2025-36911.

The attacking device — which can be a standard smartphone, tablet or laptop — broadcasts Google Fast Pair requests to any Bluetooth devices within a 14-meter radius. As it turns out, a long list of headphones from Sony, JBL, Redmi, Anker, Marshall, Jabra, OnePlus, and even Google itself (the Pixel Buds 2) will respond to these pings even when they aren’t looking to pair. On average, the attack takes just 10 seconds.

Once the headphones are paired, the attacker can do pretty much anything the owner can: listen in through the microphone, blast music, or — in some cases — locate the headset on a map if it supports Google Find Hub. That latter feature, designed strictly for finding lost headphones, creates a perfect opening for stealthy remote tracking. And here’s the twist: it’s actually most dangerous for Apple users and anyone else rocking non-Android hardware.

Remote tracking and the risks for iPhones

When headphones or a headset first shake hands with an Android device via the Fast Pair protocol, an owner key tied to that smartphone’s Google account is tucked away in the accessory’s memory. This info allows the headphones to be found later by leveraging data collected from millions of Android devices. If any random smartphone spots the target device nearby via Bluetooth, it reports its location to the Google servers. This feature — Google Find Hub — is essentially the Android version of Apple’s Find My, and it introduces the same unauthorized tracking risks as a rogue AirTag.

When an attacker hijacks the pairing, their key can be saved as the headset owner’s key — but only if the headset targeted via WhisperPair hasn’t previously been linked to an Android device and has only been used with an iPhone, or other hardware like a laptop with a different OS. Once the headphones are paired, the attacker can stalk their location on a map at their leisure — crucially, anywhere at all (not just within the 14-meter range).

Android users who’ve already used Fast Pair to link their vulnerable headsets are safe from this specific move, since they’re already logged in as the official owners. Everyone else, however, should probably double-check their manufacturer’s documentation to see if they’re in the clear — thankfully, not every device vulnerable to the exploit actually supports Google Find Hub.

How to neutralize the WhisperPair threat

The only truly effective way to fix this bug is to update your headphones’ firmware, provided an update is actually available. You can typically check for and install updates through the headset’s official companion app. The researchers have compiled a list of vulnerable devices on their site, but it’s almost certainly not exhaustive.

After updating the firmware, you absolutely must perform a factory reset to wipe the list of paired devices — including any unwanted guests.

If no firmware update is available and you’re using your headset with iOS, macOS, Windows, or Linux, your only remaining option is to track down an Android smartphone (or find a trusted friend who has one) and use it to reserve the role of the original owner. This will prevent anyone else from adding your headphones to Google Find Hub behind your back.

The update from Google

In January 2026, Google pushed an Android update to patch the vulnerability on the OS side. Unfortunately, the specifics haven’t been made public, so we’re left guessing exactly what they tweaked under the hood. Most likely, updated smartphones will no longer report the location of accessories hijacked via WhisperPair to the Google Find Hub network. But given that not everyone is exactly speedy when it comes to installing Android updates, it’s a safe bet that this type of headset tracking will remain viable for at least another couple of years.

Want to find out how else your gadgets might be spying on you? Check out these posts:

• How to protect yourself from Bluetooth stalking and more
• How to track anyone via the Find My network
• How to stop being tracked via Bluetooth beacons like AirTag
• How smartphones build a dossier on you
• Why data brokers build dossiers on you, and how to stop them doing so

Brand, website, and corporate mailout impersonation is becoming an increasingly common technique used by cybercriminals. The World Intellectual Property Organization (WIPO) reported a spike in such incidents in 2025. While tech companies and consumer brands are the most frequent targets, every industry in every country is generally at risk. The only thing that changes is how the imposters exploit the fakes In practice, we typically see the following attack scenarios:

• Luring clients and customers to a fake website to harvest login credentials for the real online store, or to steal payment details for direct theft.
• Luring employees and business partners to a fake corporate login portal to acquire legitimate credentials for infiltrating the corporate network.
• Prompting clients and customers to contact the scammers under various pretexts: getting tech support, processing a refund, entering a prize giveaway, or claiming compensation for public events involving the brand. The goal is to then swindle the victims out of as much money as possible.
• Luring business partners and employees to specially crafted pages that mimic internal company systems, to get them to approve a payment or redirect a legitimate payment to the scammers.
• Prompting clients, business partners, and employees to download malware — most often an infostealer — disguised as corporate software from a fake company website.

The words “luring” and “prompting” here imply a whole toolbox of tactics: email, messages in chat apps, social media posts that look like official ads, lookalike websites promoted through SEO tools, and even paid ads.

These schemes all share two common features. First, the attackers exploit the organization’s brand, and strive to mimic its official website, domain name, and corporate style of emails, ads, and social media posts. And the forgery doesn’t have to be flawless — just convincing enough for at least some of business partners and customers. Second, while the organization and its online resources aren’t targeted directly, the impact on them is still significant.

Business damage from brand impersonation

When fakes are crafted to target employees, an attack can lead to direct financial loss. An employee might be persuaded to transfer company funds, or their credentials could be used to steal confidential information or launch a ransomware attack.

Attacks on customers don’t typically imply direct damage to the company’s coffers, but they cause substantial indirect harm in the following areas:

• Strain on customer support. Customers who “bought” a product on a fake site will likely bring their issues to the real customer support team. Convincing them that they never actually placed an order is tough, making each case a major time waster for multiple support agents.
• Reputational damage. Defrauded customers often blame the brand for failing to protect them from the scam, and also expect compensation. According to a European survey, around half of affected buyers expect payouts and may stop using the company’s services — often sharing their negative experience on social media. This is especially damaging if the victims include public figures or anyone with a large following.
• Unplanned response costs. Depending on the specifics and scale of an attack, an affected company might need digital forensics and incident response (DFIR) services, as well as consultants specializing in consumer law, intellectual property, cybersecurity, and crisis PR.
• Increased insurance premiums. Companies that insure businesses against cyber-incidents factor in fallout from brand impersonation. An increased risk profile may be reflected in a higher premium for a business.
• Degraded website performance and rising ad costs. If criminals run paid ads using a brand’s name, they siphon traffic away from its official site. Furthermore, if a company pays to advertise its site, the cost per click rises due to the increased competition. This is a particularly acute problem for IT companies selling online services, but it’s also relevant for retail brands.
• Long-term metric decline. This includes drops in sales volume, market share, and market capitalization. These are all consequences of lost trust from customers and business partners following major incidents.

Does insurance cover the damage?

Popular cyber-risk insurance policies typically only cover costs directly tied to incidents explicitly defined in the policy — think data loss, business interruption, IT system compromise, and the like. Fake domains and web pages don’t directly damage a company’s IT systems, so they’re usually not covered by standard insurance. Reputational losses and the act of impersonation itself are separate insurance risks, requiring expanded coverage for this scenario specifically.

Of the indirect losses we’ve listed above, standard insurance might cover DFIR expenses and, in some cases, extra customer support costs (if the situation is recognized as an insured event). Voluntary customer reimbursements, lost sales, and reputational damage are almost certainly not covered.

What to do if your company is attacked by clones

If you find out someone is using your brand’s name for fraud, it makes sense to do the following:

• Send clear, straightforward notifications to your customers explaining what happened, what measures are being taken, and how to verify the authenticity of official websites, emails, and other communications.
• Create a simple “trust center” page listing your official domains, social media accounts, app store links, and support contacts. Make it easy to find and keep it updated.
• Monitor new registrations of social media pages and domain names that contain your brand names to spot the clones before an attack kicks off.
• Follow a takedown procedure. This involves gathering evidence, filing complaints with domain registrars, hosting providers, and social media administrators, then tracking the status until the fakes are fully removed. For a complete and accurate record of violations, preserve URLs, screenshots, metadata, and the date and time of discovery. Ideally, also examine the source code of fake pages, as it might contain clues pointing to other components of the criminal operation.
• Add a simple customer reporting form for suspicious sites or messages to your official website and/or branded app. This helps you learn about problems early.
• Coordinate activities between your legal, cybersecurity, and marketing teams. This ensures a consistent, unified, and effective response.

How to defend against brand impersonation attacks

While the open nature of the internet and the specifics of these attacks make preventing them outright impossible, a business can stay on top of new fakes and have the tools ready to fight back.

• Continuously monitor for suspicious public activity using specialized monitoring services. The most obvious indicator is the registration of domains similar to your brand name, but there are others — like someone buying databases related to your organization on the dark web. Comprehensive monitoring of all platforms is best outsourced to a specialized service provider, such as Kaspersky Digital Footprint Intelligence (DFI).
• The quickest and simplest way to take down a fake website or social media profile is to file a trademark infringement complaint. Make sure your portfolio of registered trademarks is robust enough to file complaints under UDRP procedures before you need it.
• When you discover fakes, deploy UDRP procedures promptly to have the fake domains transferred or removed. For social media, follow the platform’s specific infringement procedure — easily found by searching for “[social media name] trademark infringement” (for example, “LinkedIn trademark infringement”). Transferring the domain to the legitimate owner is preferred over deletion, as it prevents scammers from simply re-registering it. Many continuous monitoring services, such as Kaspersky Digital Footprint Intelligence, also offer a rapid takedown service, filing complaints on the protected brand’s behalf.
• Act quickly to block fake domains on your corporate systems. This won’t protect partners or customers, but it’ll throw a wrench into attacks targeting your own employees.
• Consider proactively registering your company’s website name and common variations (for example, with and without hyphens) in all major top-level domains, such as .com, and local extensions. This helps protect partners and customers from common typos and simple copycat sites.

The post Insider Threats: Turning 2025 Intelligence into a 2026 Defense Strategy appeared first on Flashpoint.

In 2025, cybersecurity researchers discovered several open databases belonging to various AI image-generation tools. This fact alone makes you wonder just how much AI startups care about the privacy and security of their users’ data. But the nature of the content in these databases is far more alarming.

A large number of generated pictures in these databases were images of women in lingerie or fully nude. Some were clearly created from children’s photos, or intended to make adult women appear younger (and undressed). Finally, the most disturbing part: some pornographic images were generated from completely innocent photos of real people — likely taken from social media.

In this post, we’re talking about what sextortion is, and why AI tools mean anyone can become a victim. We detail the contents of these open databases, and give you advice on how to avoid becoming a victim of AI-era sextortion.

What is sextortion?

Online sexual extortion has become so common it’s earned its own global name: sextortion (a portmanteau of sex and extortion). We’ve already detailed its various types in our post, Fifty shades of sextortion. To recap, this form of blackmail involves threatening to publish intimate images or videos to coerce the victim into taking certain actions, or to extort money from them.

Previously, victims of sextortion were typically adult industry workers, or individuals who’d shared intimate content with an untrustworthy person.

However, the rapid advancement of artificial intelligence, particularly text-to-image technology, has fundamentally changed the game. Now, literally anyone who’s posted their most innocent photos publicly can become a victim of sextortion. This is because generative AI makes it possible to quickly, easily, and convincingly undress people in any digital image, or add a generated nude body to someone’s head in a matter of seconds.

Of course, this kind of fakery was possible before AI, but it required long hours of meticulous Photoshop work. Now, all you need is to describe the desired result in words.

To make matters worse, many generative AI services don’t bother much with protecting the content they’ve been used to create. As mentioned earlier, last year saw researchers discover at least three publicly accessible databases belonging to these services. This means the generated nudes within them were available not just to the user who’d created them, but to anyone on the internet.

How the AI image database leak was discovered

In October 2025, cybersecurity researcher Jeremiah Fowler uncovered an open database containing over a million AI-generated images and videos. According to the researcher, the overwhelming majority of this content was pornographic in nature. The database wasn’t encrypted or password-protected — meaning any internet user could access it.

The database’s name and watermarks on some images led Fowler to believe its source was the U.S.-based company SocialBook, which offers services for influencers and digital marketing services. The company’s website also provides access to tools for generating images and content using AI.

However, further analysis revealed that SocialBook itself wasn’t directly generating this content. Links within the service’s interface led to third-party products — the AI services MagicEdit and DreamPal — which were the tools used to create the images. These tools allowed users to generate pictures from text descriptions, edit uploaded photos, and perform various visual manipulations, including creating explicit content and face-swapping.

The leak was linked to these specific tools, and the database contained the product of their work, including AI-generated and AI-edited images. A portion of the images led the researcher to suspect they’d been uploaded to the AI as references for creating provocative imagery.

Fowler states that roughly 10,000 photos were being added to the database every single day. SocialBook denies any connection to the database. After the researcher informed the company of the leak, several pages on the SocialBook website that had previously mentioned MagicEdit and DreamPal became inaccessible and began returning errors.

Which services were the source of the leak?

Both services — MagicEdit and DreamPal — were initially marketed as tools for interactive, user-driven visual experimentation with images and art characters. Unfortunately, a significant portion of these capabilities were directly linked to creating sexualized content.

For example, MagicEdit offered a tool for AI-powered virtual clothing changes, as well as a set of styles that made images of women more revealing after processing — such as replacing everyday clothes with swimwear or lingerie. Its promotional materials promised to turn an ordinary look into a sexy one in seconds.

DreamPal, for its part, was initially positioned as an AI-powered role-playing chat, and was even more explicit about its adult-oriented positioning. The site offered to create an ideal AI girlfriend, with certain pages directly referencing erotic content. The FAQ also noted that filters for explicit content in chats were disabled so as not to limit users’ most intimate fantasies.

Both services have suspended operations. At the time of writing, the DreamPal website returned an error, while MagicEdit seemed available again. Their apps were removed from both the App Store and Google Play.

Jeremiah Fowler says earlier in 2025, he discovered two more open databases containing AI-generated images. One belonged to the South Korean site GenNomis, and contained 95,000 entries — a substantial portion of which being images of “undressed” people. Among other things, the database included images with child versions of celebrities: American singers Ariana Grande and Beyoncé, and reality TV star Kim Kardashian.

How to avoid becoming a victim

In light of incidents like these, it’s clear that the risks associated with sextortion are no longer confined to private messaging or the exchange of intimate content. In the era of generative AI, even ordinary photos, when posted publicly, can be used to create compromising content.

This problem is especially relevant for women, but men shouldn’t get too comfortable either: the popular blackmail scheme of “I hacked your computer and used the webcam to make videos of you browsing adult sites” could reach a whole new level of persuasion thanks to AI tools for generating photos and videos.

Therefore, protecting your privacy on social media and controlling what data about you is publicly available become key measures for safeguarding both your reputation and peace of mind. To prevent your photos from being used to create questionable AI-generated content, we recommend making all your social media profiles as private as possible — after all, they could be the source of images for AI-generated nudes.

We’ve already published multiple detailed guides on how to reduce your digital footprint online or even remove your data from the internet, how to stop data brokers from compiling dossiers on you, and protect yourself from intimate image abuse.

Additionally, we have a dedicated service, Privacy Checker — perfect for anyone who wants a quick but systematic approach to privacy settings everywhere possible. It compiles step-by-step guides for securing accounts on social media and online services across all major platforms.

And to ensure the safety and privacy of your child’s data, Kaspersky Safe Kids can help: it allows parents to monitor which social media their child spends time on. From there, you can help them adjust privacy settings on their accounts so their posted photos aren’t used to create inappropriate content. Explore our guide to children’s online safety together, and if your child dreams of becoming a popular blogger, discuss our step-by-step cybersecurity guide for wannabe bloggers with them.

Database platform MongoDB disclosed CVE-2025-14847, called MongoBleed. This is an unauthenticated memory disclosure vulnerability with a CVSS score of 8.7.

The post Threat Brief: MongoDB Vulnerability (CVE-2025-14847) appeared first on Unit 42.

Thanks to the convenience of NFC and smartphone payments, many people no longer carry wallets or remember their bank card PINs. All their cards reside in a payment app, and using that is quicker than fumbling for a physical card. Mobile payments are also secure — the technology was developed relatively recently and includes numerous anti-fraud protections. Still, criminals have invented several ways to abuse NFC and steal your money. Fortunately, protecting your funds is straightforward: just know about these tricks and avoid risky NFC usage scenarios.

What are NFC relay and NFCGate?

NFC relay is a technique where data wirelessly transmitted between a source (like a bank card) and a receiver (like a payment terminal) is intercepted by one intermediate device, and relayed in real time to another. Imagine you have two smartphones connected via the internet, each with a relay app installed. If you tap a physical bank card against the first smartphone and hold the second smartphone near a terminal or ATM, the relay app on the first smartphone will read the card’s signal using NFC, and relay it in real time to the second smartphone, which will then transmit this signal to the terminal. From the terminal’s perspective, it all looks like a real card is tapped on it — even though the card itself might physically be in another city or country.

This technology wasn’t originally created for crime. The NFCGate app appeared in 2015 as a research tool after it was developed by students at the Technical University of Darmstadt in Germany. It was intended for analyzing and debugging NFC traffic, as well as for education purposes and experiments with contactless technology. NFCGate was distributed as an open-source solution and used in academic and enthusiast circles.

Five years later, cybercriminals caught on to the potential of NFC relay and began modifying NFCGate by adding mods that allowed it to run through a malicious server, disguise itself as legitimate software, and perform social engineering scenarios.

What began as a research project morphed into the foundation for an entire class of attacks aimed at draining bank accounts without physical access to bank cards.

A history of misuse

The first documented attacks using a modified NFCGate occurred in late 2023 in the Czech Republic. By early 2025, the problem had become large scale  and noticeable: cybersecurity analysts uncovered more than 80 unique malware samples built on the NFCGate framework. The attacks evolved rapidly, with NFC relay capabilities being integrated into other malware components.

By February 2025, malware bundles combining CraxsRAT and NFCGate emerged, allowing attackers to install and configure the relay with minimal victim interaction. A new scheme, a so-called “reverse” version of NFCGate, appeared in spring 2025, fundamentally changing the attack’s execution.

Particularly noteworthy is the RatOn Trojan, first detected in the Czech Republic. It combines remote smartphone control with NFC relay capabilities, letting attackers target victims’ banking apps and cards through various technique combinations. Features like screen capture, clipboard data manipulation, SMS sending, and stealing info from crypto wallets and banking apps give criminals an extensive arsenal.

Cybercriminals have also packaged NFC relay technology into malware-as-a-service (MaaS) offerings, and reselling them to other threat actors through subscription. In early 2025, analysts uncovered a new and sophisticated Android malware campaign in Italy, dubbed SuperCard X. Attempts to deploy SuperCard X were recorded in Russia in May 2025, and in Brazil in August of the same year.

The direct NFCGate attack

The direct attack is the original criminal scheme exploiting NFCGate. In this scenario, the victim’s smartphone plays the role of the reader, while the attacker’s phone acts as the card emulator.

First, the fraudsters trick the user into installing a malicious app disguised as a banking service, a system update, an “account security” app, or even a popular app like TikTok. Once installed, the app gains access to both NFC and the internet — often without requesting dangerous permissions or root access. Some versions also ask for access to Android accessibility features.

Then, under the guise of identity verification, the victim is prompted to tap their bank card to their phone. When they do, the malware reads the card data via NFC and immediately sends it to the criminals’ server. From there, the information is relayed to a second smartphone held by a money mule, who helps extract the money. This phone then emulates the victim’s card to make payments at a terminal or withdraw cash from an ATM.

The fake app on the victim’s smartphone also asks for the card PIN — just like at a payment terminal or ATM — and sends it to the attackers.

In early versions of the attack, criminals would simply stand ready at an ATM with a phone to use the duped user’s card in real time. Later, the malware was refined so the stolen data could be used for in-store purchases in a delayed, offline mode, rather than in a live relay.

For the victim, the theft is hard to notice: the card never left their possession, they didn’t have to manually enter or recite its details, and the bank alerts about the withdrawals can be delayed or even intercepted by the malicious app itself.

Among the red flags that should make you suspect a direct NFC attack are:

• prompts to install apps not from official stores;
• requests to tap your bank card on your phone.

The reverse NFCGate attack

The reverse attack is a newer, more sophisticated scheme. The victim’s smartphone no longer reads their card — it emulates the attacker’s card. To the victim, everything appears completely safe: there’s no need to recite card details, share codes, or tap a card to the phone.

Just like with the direct scheme, it all starts with social engineering. The user gets a call or message convincing them to install an app for “contactless payments”, “card security”, or even “using central bank digital currency”. Once installed, the new app asks to be set as the default contactless payment method — and this step is critically important. Thanks to this, the malware requires no root access — just user consent.

The malicious app then silently connects to the attackers’ server in the background, and the NFC data from a card belonging to one of the criminals is transmitted to the victim’s device. This step is completely invisible to the victim.

Next, the victim is directed to an ATM. Under the pretext of “transferring money to a secure account” or “sending money to themselves”, they are instructed to tap their phone on the ATM’s NFC reader. At this moment, the ATM is actually interacting with the attacker’s card. The PIN is dictated to the victim beforehand — presented as “new” or “temporary”.

The result is that all the money deposited or transferred by the victim ends up in the criminals’ account.

The hallmarks of this attack are:

• requests to change your default NFC payment method;
• a “new” PIN;
• any scenario where you’re told to go to an ATM and perform actions there under someone else’s instructions.

How to protect yourself from NFC relay attacks

NFC relay attacks rely not so much on technical vulnerabilities as on user trust. Defending against them comes down to some simple precautions.

• Make sure you keep your trusted contactless payment method (like Google Pay or Samsung Pay) as the default.
• Never tap your bank card on your phone at someone else’s request, or because an app tells you to. Legitimate apps might use your camera to scan a card number, but they’ll never ask you to use the NFC reader for your own card.
• Never follow instructions from strangers at an ATM — no matter who they claim to be.
• Avoid installing apps from unofficial sources. This includes links sent via messaging apps, social media, SMS, or recommended during a phone call — even if they come from someone claiming to be customer support or the police.
• Use comprehensive security on your Android smartphones to block scam calls, prevent visits to phishing sites, and stop malware installation.
• Stick to official app stores only. When downloading from a store, check the app’s reviews, number of downloads, publication date, and rating.
• When using an ATM, rely on your physical card instead of your smartphone for the transaction.
• Make it a habit to regularly check the “Payment default” setting in your phone’s NFC menu. If you see any suspicious apps listed, remove them immediately and run a full security scan on your device.
• Review the list of apps with accessibility permissions — this is a feature commonly abused by malware. Either revoke these permissions for any suspicious apps, or uninstall the apps completely.
• Save the official customer service numbers for your banks in your phone’s contacts. At the slightest hint of foul play, call your bank’s hotline directly without delay.
• If you suspect your card details may have been compromised, block the card immediately.

Statistics across all threats

In Q3 2025, the percentage of ICS computers on which malicious objects were blocked decreased from the previous quarter by 0.4 pp to 20.1%. This is the lowest level for the observed period.

Regionally, the percentage of ICS computers on which malicious objects were blocked ranged from 9.2% in Northern Europe to 27.4% in Africa.

In Q3 2025, the percentage increased in five regions. The most notable increase occurred in East Asia, triggered by the local spread of malicious scripts in the OT infrastructure of engineering organizations and ICS integrators.

Selected industries

The biometrics sector traditionally led the rankings of the industries and OT infrastructures surveyed in this report in terms of the percentage of ICS computers on which malicious objects were blocked.

In Q3 2025, the percentage of ICS computers on which malicious objects were blocked increased in four of the seven surveyed industries. The most notable increases were in engineering and ICS integrators, and manufacturing.

Diversity of detected malicious objects

In Q3 2025, Kaspersky protection solutions blocked malware from 11,356 different malware families of various categories on industrial automation systems.

In Q3 2025, there was a decrease in the percentage of ICS computers on which denylisted internet resources and miners of both categories were blocked. These were the only categories that exhibited a decrease.

Main threat sources

Depending on the threat detection and blocking scenario, it is not always possible to reliably identify the source. The circumstantial evidence for a specific source can be the blocked threat’s type (category).

The internet (visiting malicious or compromised internet resources; malicious content distributed via messengers; cloud data storage and processing services and CDNs), email clients (phishing emails), and removable storage devices remain the primary sources of threats to computers in an organization’s technology infrastructure.

In Q3 2025, the percentage of ICS computers on which malicious objects from various sources were blocked decreased.

The same computer can be attacked by several categories of malware from the same source during a quarter. That computer is counted when calculating the percentage of attacked computers for each threat category, but is only counted once for the threat source (we count unique attacked computers). In addition, it is not always possible to accurately determine the initial infection attempt. Therefore, the total percentage of ICS computers on which various categories of threats from a certain source were blocked can exceed the percentage of threats from the source itself.

• The main categories of threats from the internet blocked on ICS computers in Q3 2025 were malicious scripts and phishing pages, and denylisted internet resources. The percentage ranged from 4.57% in Northern Europe to 10.31% in Africa.
• The main categories of threats from email clients blocked on ICS computers were malicious scripts and phishing pages, spyware, and malicious documents. Most of the spyware detected in phishing emails was delivered as a password-protected archive or a multi-layered script embedded in an office document. The percentage of ICS computers on which threats from email clients were blocked ranged from 0.78% in Russia to 6.85% in Southern Europe.
• The main categories of threats that were blocked when removable media was connected to ICS computers were worms, viruses, and spyware. The percentage of ICS computers on which threats from this source were blocked ranged from 0.05% in Australia and New Zealand to 1.43% in Africa.
• The main categories of threats that spread through network folders were viruses, AutoCAD malware, worms, and spyware. The percentages of ICS computers where threats from this source were blocked ranged from 0.006% in Northern Europe to 0.20% in East Asia.

Threat categories

Typical attacks blocked within an OT network are multi-step sequences of malicious activities, where each subsequent step of the attackers is aimed at increasing privileges and/or gaining access to other systems by exploiting the security problems of industrial enterprises, including technological infrastructures.

Malicious objects used for initial infection

In Q3 2025, the percentage of ICS computers on which denylisted internet resources were blocked decreased to 4.01%. This is the lowest quarterly figure since the beginning of 2022.

Regionally, the percentage of ICS computers on which denylisted internet resources were blocked ranged from 2.35% in Australia and New Zealand to 4.96% in Africa. Southeast Asia and South Asia were also among the top three regions for this indicator.

The percentage of ICS computers on which malicious documents were blocked has grown for three consecutive quarters, following a decline at the end of 2024. In Q3 2025, it reached 1,98%.

The indicator increased in four regions: South America, East Asia, Southeast Asia, and Australia and New Zealand. South America saw the largest increase as a result of a large-scale phishing campaign in which attackers used new exploits for an old vulnerability (CVE-2017-11882) in Microsoft Office Equation Editor to deliver various spyware to victims’ computers. It is noteworthy that the attackers in this phishing campaign used localized Spanish-language emails disguised as business correspondence.

In Q3 2025, the percentage of ICS computers on which malicious scripts and phishing pages were blocked increased to 6.79%. This category led the rankings of threat categories in terms of the percentage of ICS computers on which they were blocked.

Regionally, the percentage of ICS computers on which malicious scripts and phishing pages were blocked ranged from 2.57% in Northern Europe to 9.41% in Africa. The top three regions for this indicator were Africa, East Asia, and South America. The indicator increased the most in East Asia (by a dramatic 5.23 pp) as a result of the local spread of malicious spyware scripts loaded into the memory of popular torrent clients including MediaGet.

Next-stage malware

Malicious objects used to initially infect computers deliver next-stage malware — spyware, ransomware, and miners — to victims’ computers. As a rule, the higher the percentage of ICS computers on which the initial infection malware is blocked, the higher the percentage for next-stage malware.
In Q3 2025, the percentage of ICS computers on which spyware and ransomware were blocked increased. The rates were:

• spyware: 4.04% (up 0.20 pp);
• ransomware: 0.17% (up 0.03 pp).

The percentage of ICS computers on which miners of both categories were blocked decreased. The rates were:

• miners in the form of executable files for Windows: 0.57% (down 0.06 pp), it’s the lowest level since Q3 2022;
• web miners: 0.25% (down 0.05 pp). This is the lowest level since Q3 2022.

Self-propagating malware

Self-propagating malware (worms and viruses) is a category unto itself. Worms and virus-infected files were originally used for initial infection, but as botnet functionality evolved, they took on next-stage characteristics.

To spread across ICS networks, viruses and worms rely on removable media and network folders in the form of infected files, such as archives with backups, office documents, pirated games and hacked applications. In rarer and more dangerous cases, web pages with network equipment settings, as well as files stored in internal document management systems, product lifecycle management (PLM) systems, resource management (ERP) systems and other web services are infected.

In Q3 2025, the percentage of ICS computers on which worms and viruses were blocked increased to 1.26% (by 0.04 pp) and 1.40% (by 0.11 pp), respectively.

AutoCAD malware

This category of malware can spread in a variety of ways, so it does not belong to a specific group.

In Q3 2025, the percentage of ICS computers on which AutoCAD malware was blocked slightly increased to 0.30% (by 0.01 pp).

For more information on industrial threats see the full version of the report.

Introduction

If you’re a penetration tester, you know that lateral movement is becoming increasingly difficult, especially in well-defended environments. One common technique for remote command execution has been the use of DCOM objects.

Over the years, many different DCOM objects have been discovered. Some rely on native Windows components, others depend on third-party software such as Microsoft Office, and some are undocumented objects found through reverse engineering. While certain objects still work, others no longer function in newer versions of Windows.

This research presents a previously undescribed DCOM object that can be used for both command execution and potential persistence. This new technique abuses older initial access and persistence methods through Control Panel items.

First, we will discuss COM technology. After that, we will review the current state of the Impacket dcomexec script, focusing on objects that still function, and discuss potential fixes and improvements, then move on to techniques for enumerating objects on the system. Next, we will examine Control Panel items, how adversaries have used them for initial access and persistence, and how these items can be leveraged through a DCOM object to achieve command execution.

Finally, we will cover detection strategies to identify and respond to this type of activity.

COM/DCOM technology

What is COM?

COM stands for Component Object Model, a Microsoft technology that defines a binary standard for interoperability. It enables the creation of reusable software components that can interact at runtime without the need to compile COM libraries directly into an application.

These software components operate in a client–server model. A COM object exposes its functionality through one or more interfaces. An interface is essentially a collection of related member functions (methods).

COM also enables communication between processes running on the same machine by using local RPC (Remote Procedure Call) to handle cross-process communication.

Terms

To ensure a better understanding of its structure and functionality, let’s revise COM-related terminology.

1. COM interface
   A COM interface defines the functionality that a COM object exposes. Each COM interface is identified by a unique GUID known as the IID (Interface ID). All COM interfaces can be found in the Windows Registry under HKEY_CLASSES_ROOT\Interface, where they are organized by GUID.
2. COM class (COM CoClass)
   A COM class is the actual implementation of one or more COM interfaces. Like COM interfaces, classes are identified by unique GUIDs, but in this case the GUID is called the CLSID (Class ID). This GUID is used to locate the COM server and activate the corresponding COM class.

   All COM classes must be registered in the registry under HKEY_CLASSES_ROOT\CLSID, where each class’s GUID is stored. Under each GUID, you may find multiple subkeys that serve different purposes, such as:

   • InprocServer32/LocalServer32: Specifies the system path of the COM server where the class is defined. InprocServer32 is used for in-process servers (DLLs), while LocalServer32 is used for out-of-process servers (EXEs). We’ll describe this in more detail later.
   • ProgID: A human-readable name assigned to the COM class.
   • TypeLib: A binary description of the COM class (essentially documentation for the class).
   • AppID: Used to describe security configuration for the class.
3. COM server
   A COM is the module where a COM class is defined. The server can be implemented as an EXE, in which case it is called an out-of-process server, or as a DLL, in which case it is called an in-process server. Each COM server has a unique file path or location in the system. Information about COM servers is stored in the Windows Registry. The COM runtime uses the registry to locate the server and perform further actions. Registry entries for COM servers are located under the HKEY_CLASSES_ROOT root key for both 32- and 64-bit servers.

Client–server model

1. In-process server
   In the case of an in-process server, the server is implemented as a DLL. The client loads this DLL into its own address space and directly executes functions exposed by the COM object. This approach is efficient since both client and server run within the same process.
   

   In-process COM server

2. Out-of-process server
   Here, the server is implemented and compiled as an executable (EXE). Since the client cannot load an EXE into its address space, the server runs in its own process, separate from the client. Communication between the two processes is handled via ALPC (Advanced Local Procedure Call) ports, which serve as the RPC transport layer for COM.

What is DCOM?

DCOM is an extension of COM where the D stands for Distributed. It enables the client and server to reside on different machines. From the user’s perspective, there is no difference: DCOM provides an abstraction layer that makes both the client and the server appear as if they are on the same machine.

Under the hood, however, COM uses TCP as the RPC transport layer to enable communication across machines.

Certain requirements must be met to extend a COM object into a DCOM object. The most important one for our research is the presence of the AppID subkey in the registry, located under the COM CLSID entry.

The AppID value contains a GUID that maps to a corresponding key under HKEY_CLASSES_ROOT\AppID. Several subkeys may exist under this GUID. Two critical ones are:

• AccessPermission: controls access permissions.
• LaunchPermission: controls activation permissions.

These registry settings grant remote clients permissions to activate and interact with DCOM objects.

Lateral movement via DCOM

After attackers compromise a host, their next objective is often to compromise additional machines. This is what we call lateral movement. One common lateral movement technique is to achieve remote command execution on a target machine. There are many ways to do this, one of which involves abusing DCOM objects.

In recent years, many DCOM objects have been discovered. This research focuses on the objects exposed by the Impacket script dcomexec.py that can be used for command execution. More specifically, three exposed objects are used: ShellWindows, ShellBrowserWindow and MMC20.

1. ShellWindows
   ShellWindows was one of the first DCOM objects to be identified. It represents a collection of open shell windows and is hosted by explorer.exe, meaning any COM client communicates with that process.

   In Impacket’s dcomexec.py, once an instance of this COM object is created on a remote machine, the script provides a semi-interactive shell.

   Each time a user enters a command, the function exposed by the COM object is called. The command output is redirected to a file, which the script retrieves via SMB and displays back to simulate a regular shell.

   Internally, the script runs this command when connecting:

   cmd.exe /Q /c cd \ 1> \\127.0.0.1\ADMIN$\__17602 2>&1

   This sets the working directory to C:\ and redirects the output to the ADMIN$ share under the filename __17602. After that, the script checks whether the file exists; if it does, execution is considered successful and the output appears as if in a shell.

   When running dcomexec.py against Windows 10 and 11 using the ShellWindows object, the script hangs after confirming SMB connection initialization and printing the SMB banner. As I mentioned in my personal blog post, it appears that this DCOM object no longer has permission to write to the ADMIN$ share. A simple fix is to redirect the output to a directory the DCOM object can write to, such as the Temp folder. The Temp folder can then be accessed under the same ADMIN$ share. A small change in the code resolves the issue. For example:

   OUTPUT_FILENAME = 'Temp\\__' + str(time.time())[:5]

2. ShellBrowserWindow
   The ShellBrowserWindow object behaves almost identically to ShellWindows and exhibits the same behavior on Windows 10. The same workaround that we used for ShellWindows applies in this case. However, on Windows 11, this object no longer works for command execution.
3. MMC20
   The MMC20.Application COM object is the automation interface for Microsoft Management Console (MMC). It exposes methods and properties that allow MMC snap-ins to be automated.

   This object has historically worked across all Windows versions. Starting with Windows Server 2025, however, attempting to use it triggers a Defender alert, and execution is blocked.

   As shown in earlier examples, the dcomexec.py script writes the command output to a file under ADMIN$, with a filename that begins with __:

   OUTPUT_FILENAME = '__' + str(time.time())[:5]

   Defender appears to check for files written under ADMIN$ that start with __, and when it detects one, it blocks the process and alerts the user. A quick fix is to simply remove the double underscores from the output filename.

   Another way to bypass this issue is to use the same workaround used for ShellWindows – redirecting the output to the Temp folder. The table below outlines the status of these objects across different Windows versions.

   	Windows Server 2025	Windows Server 2022	Windows 11	Windows 10
   ShellWindows	Doesn’t work	Doesn’t work	Works but needs a fix	Works but needs a fix
   ShellBrowserWindow	Doesn’t work	Doesn’t work	Doesn’t work	Works but needs a fix
   MMC20	Detected by Defender	Works	Works	Works

Enumerating COM/DCOM objects

The first step to identifying which DCOM objects could be used for lateral movement is to enumerate them. By enumerating, I don’t just mean listing the objects. Enumeration involves:

• Finding objects and filtering specifically for DCOM objects.
• Identifying their interfaces.
• Inspecting the exposed functions.

Automating enumeration is difficult because most COM objects lack a type library (TypeLib). A TypeLib acts as documentation for an object: which interfaces it supports, which functions are exposed, and the definitions of those functions. Even when TypeLibs are available, manual inspection is often still required, as we will explain later.

There are several approaches to enumerating COM objects depending on their use cases. Next, we’ll describe the methods I used while conducting this research, taking into account both automated and manual methods.

1. Automation using PowerShell
   In PowerShell, you can use .NET to create and interact with DCOM objects. Objects can be created using either their ProgID or CLSID, after which you can call their functions (as shown in the figure below).
   

   Shell.Application COM object function list in PowerShell

   Under the hood, PowerShell checks whether the COM object has a TypeLib and implements the IDispatch interface. IDispatch enables late binding, which allows runtime dynamic object creation and function invocation. With these two conditions met, PowerShell can dynamically interact with COM objects at runtime.

   Our strategy looks like this:

   

   As you can see in the last box, we perform manual inspection to look for functions with names that could be of interest, such as Execute, Exec, Shell, etc. These names often indicate potential command execution capabilities.

   However, this approach has several limitations:

   • TypeLib requirement: Not all COM objects have a TypeLib, so many objects cannot be enumerated this way.
   • IDispatch requirement: Not all COM objects implement the IDispatch interface, which is required for PowerShell interaction.
   • Interface control: When you instantiate an object in PowerShell, you cannot choose which interface the instance will be tied to. If a COM class implements multiple interfaces, PowerShell will automatically select the one marked as [default] in the TypeLib. This means that other non-default interfaces, which may contain additional relevant functionality, such as command execution, could be overlooked.
2. Automation using C++
   As you might expect, C++ is one of the languages that natively supports COM clients. Using C++, you can create instances of COM objects and call their functions via header files that define the interfaces.However, with this approach, we are not necessarily interested in calling functions directly. Instead, the goal is to check whether a specific COM object supports certain interfaces. The reasoning is that many interfaces have been found to contain functions that can be abused for command execution or other purposes.

   This strategy primarily relies on an interface called IUnknown. All COM interfaces should inherit from this interface, and all COM classes should implement it.The IUnknown interface exposes three main functions. The most important is QueryInterface(), which is used to ask a COM object for a pointer to one of its interfaces.So, the strategy is to:

   • Enumerate COM classes in the system by reading CLSIDs under the HKEY_CLASSES_ROOT\CLSID key.
   • Check whether they support any known valuable interfaces. If they do, those classes may be leveraged for command execution or other useful functionality.

   This method has several advantages:

   • No TypeLib dependency: Unlike PowerShell, this approach does not require the COM object to have a TypeLib.
   • Use of IUnknown: In C++, you can use the QueryInterface function from the base IUnknown interface to check if a particular interface is supported by a COM class.
   • No need for interface definitions: Even without knowing the exact interface structure, you can obtain a pointer to its virtual function table (vtable), typically cast as a void*. This is enough to confirm the existence of the interface and potentially inspect it further.

   The figure below illustrates this strategy:

   

   This approach is good in terms of automation because it eliminates the need for manual inspection. However, we are still only checking well-known interfaces commonly used for lateral movement, while potentially missing others.

3. Manual inspection using open-source tools
   As you can see, automation can be difficult since it requires several prerequisites and, in many cases, still ends with a manual inspection. An alternative approach is manual inspection using a tool called OleViewDotNet, developed by James Forshaw. This tool allows you to:
   • List all COM classes in the system.
   • Create instances of those classes.
   • Check their supported interfaces.
   • Call specific functions.
   • Apply various filters for easier analysis.
   • Perform other inspection tasks.
   

   Open-source tool for inspecting COM interfaces

   One of the most valuable features of this tool is its naming visibility. OleViewDotNet extracts the names of interfaces and classes (when available) from the Windows Registry and displays them, along with any associated type libraries.

   This makes manual inspection easier, since you can analyze the names of classes, interfaces, or type libraries and correlate them with potentially interesting functionality, for example, functions that could lead to command execution or persistence techniques.

Control Panel items as attack surfaces

Control Panel items allow users to view and adjust their computer settings. These items are implemented as DLLs that export the CPlApplet function and typically have the .cpl extension. Control Panel items can also be executables, but our research will focus on DLLs only.

Attackers can abuse CPL files for initial access. When a user executes a malicious .cpl file (e.g., delivered via phishing), the system may be compromised – a technique mapped to MITRE ATT&CK T1218.002.

Adversaries may also modify the extensions of malicious DLLs to .cpl and register them in the corresponding locations in the registry.

• Under HKEY_CURRENT_USER:
  HKCU\Software\Microsoft\Windows\CurrentVersion\Control Panel\Cpls

• Under HKEY_LOCAL_MACHINE:
  • For 64-bit DLLs:
    HKLM\Software\Microsoft\Windows\CurrentVersion\Control Panel\Cpls
  • For 32-bit DLLs:
    HKLM\Software\WOW6432Node\Microsoft\Windows\CurrentVersion\Control Panel\Cpls

These locations are important when Control Panel DLLs need to be available to the current logged-in user or to all users on the machine. However, the “Control Panel” subkey and its “Cpls” subkey under HKCU should be created manually, unlike the “Control Panel” and “Cpls” subkeys under HKLM, which are created automatically by the operating system.

Once registered, the DLL (CPL file) will load every time the Control Panel is opened, enabling persistence on the victim’s system.

It’s worth noting that even DLLs that do not comply with the CPL specification, do not export CPlApplet, or do not have the .cpl extension can still be executed via their DllEntryPoint function if they are registered under the registry keys listed above.

There are multiple ways to execute Control Panel items:

• From cmd: control.exe [filename].cpl
• By double-clicking the .cpl file.

Both methods use rundll32.exe under the hood:

rundll32.exe shell32.dll,Control_RunDLL [filename].cpl

This calls the Control_RunDLL function from shell32.dll, passing the CPL file as an argument. Everything inside the CPlApplet function will then be executed.

However, if the CPL file has been registered in the registry as shown earlier, then every time the Control Panel is opened, the file is loaded into memory through the COM Surrogate process (dllhost.exe):

What happened was that a Control Panel with a COM client used a COM object to load these CPL files. We will talk about this COM object in more detail later.

The COM Surrogate process was designed to host COM server DLLs in a separate process rather than loading them directly into the client process’s address space. This isolation improves stability for the in-process server model. This hosting behavior can be configured for a COM object in the registry if you want a COM server DLL to run inside a separate process because, by default, it is loaded in the same process.

‘DCOMing’ through Control Panel items

While following the manual approach of enumerating COM/DCOM objects that could be useful for lateral movement, I came across a COM object called COpenControlPanel, which is exposed through shell32.dll and has the CLSID {06622D85-6856-4460-8DE1-A81921B41C4B}. This object exposes multiple interfaces, one of which is IOpenControlPanel with IID {D11AD862-66DE-4DF4-BF6C-1F5621996AF1}.

I immediately thought of its potential to compromise Control Panel items, so I wanted to check which functions were exposed by this interface. Unfortunately, neither the interface nor the COM class has a type library.

Normally, checking the interface definition would require reverse engineering, so at first, it looked like we needed to take a different research path. However, it turned out that the IOpenControlPanel interface is documented on MSDN, and according to the documentation, it exposes several functions. One of them, called Open, allows a specified Control Panel item to be opened using its name as the first argument.

Full type and function definitions are provided in the shobjidl_core.h Windows header file.

It’s worth noting that in newer versions of Windows (e.g., Windows Server 2025 and Windows 11), Microsoft has removed interface names from the registry, which means they can no longer be identified through OleViewDotNet.

Returning to the COpenControlPanel COM object, I found that the Open function can trigger a DLL to be loaded into memory if it has been registered in the registry. For the purposes of this research, I created a DLL that basically just spawns a message box which is defined under the DllEntryPoint function. I registered it under HKCU\Software\Microsoft\Windows\CurrentVersion\Control Panel\Cpls and then created a simple C++ COM client to call the Open function on this interface.

As expected, the DLL was loaded into memory. It was hosted in the same way that it would be if the Control Panel itself was opened: through the COM Surrogate process (dllhost.exe). Using Process Explorer, it was clear that dllhost.exe loaded my DLL while simultaneously hosting the COpenControlPanel object along with other COM objects.

Based on my testing, I made the following observations:

1. The DLL that needs to be registered does not necessarily have to be a .cpl file; any DLL with a valid entry point will be loaded.
2. The Open() function accepts the name of a Control Panel item as its first argument. However, it appears that even if a random string is supplied, it still causes all DLLs registered in the relevant registry location to be loaded into memory.

Now, what if we could trigger this COM object remotely? In other words, what if it is not just a COM object but also a DCOM object? To verify this, we checked the AppID of the COpenControlPanel object using OleViewDotNet.

Both the launch and access permissions are empty, which means the object will follow the system’s default DCOM security policy. By default, members of the Administrators group are allowed to launch and access the DCOM object.

Based on this, we can build a remote strategy. First, upload the “malicious” DLL, then use the Remote Registry service to register it in the appropriate registry location. Finally, use a trigger acting as a DCOM client to remotely invoke the Open() function, causing our DLL to be loaded. The diagram below illustrates the flow of this approach.

The trigger can be written in either C++ or Python, for example, using Impacket. I chose Python because of its flexibility. The trigger itself is straightforward: we define the DCOM class, the interface, and the function to call. The full code example can be found here.

Once the trigger runs, the behavior will be the same as when executing the COM client locally: our DLL will be loaded through the COM Surrogate process (dllhost.exe).

As you can see, this technique not only achieves command execution but also provides persistence. It can be triggered in two ways: when a user opens the Control Panel or remotely at any time via DCOM.

Detection

The first step in detecting such activity is to check whether any Control Panel items have been registered under the following registry paths:

• HKCU\Software\Microsoft\Windows\CurrentVersion\Control Panel\Cpls
• HKLM\Software\Microsoft\Windows\CurrentVersion\Control Panel\Cpls
• HKLM\Software\WOW6432Node\Microsoft\Windows\CurrentVersion\Control Panel\Cpls

Although commonly known best practices and research papers regarding Windows security advise monitoring only the first subkey, for thorough coverage it is important to monitor all of the above.

In addition, monitoring dllhost.exe (COM Surrogate) for unusual COM objects such as COpenControlPanel can provide indicators of malicious activity.
Finally, it is always recommended to monitor Remote Registry usage because it is commonly abused in many types of attacks, not just in this scenario.

Conclusion

In conclusion, I hope this research has clarified yet another attack vector and emphasized the importance of implementing hardening practices. Below are a few closing points for security researchers to take into account:

• As shown, DCOM represents a large attack surface. Windows exposes many DCOM classes, a significant number of which lack type libraries – meaning reverse engineering can reveal additional classes that may be abused for lateral movement.
• Changing registry values to register malicious CPLs is not good practice from a red teaming ethics perspective. Defender products tend to monitor common persistence paths, but Control Panel applets can be registered in multiple registry locations, so there is always a gap that can be exploited.
• Bitness also matters. On x64 systems, loading a 32-bit DLL will spawn a 32-bit COM Surrogate process (dllhost.exe *32). This is unusual on 64-bit hosts and therefore serves as a useful detection signal for defenders and an interesting red flag for red teamers to consider.

In November 2025, Kaspersky experts uncovered a new stealer named Stealka, which targets Windows users’ data. Attackers are using Stealka to hijack accounts, steal cryptocurrency, and install a crypto miner on their victims’ devices. Most frequently, this infostealer disguises itself as game cracks, cheats and mods.

Here’s how the attackers are spreading the stealer, and how you can protect yourself.

How Stealka spreads

A stealer is a type of malware that collects confidential information stored on the victim’s device and sends it to the attackers’ server. Stealka is primarily distributed via popular platforms like GitHub, SourceForge, Softpedia, sites.google.com, and others, disguised as cracks for popular software, or cheats and mods for games. For the malware to be activated, the user must run the file manually.

Here’s an example: a malicious Roblox mod published on SourceForge.

And here’s one on GitHub posing as a crack for Microsoft Visio.

Sometimes, however, attackers go a step further (and possibly use AI tools) to create entire fake websites that look quite professional. Without the help of a robust antivirus, the average user is unlikely to realize anything is amiss.

Admittedly, the cracks and software advertised on these fake sites can sometimes look a bit off. For example, here the attackers are offering a download for Half-Life 3, while at the same time claiming it’s not actually a game but some kind of “professional software solution designed for Windows”.

The truth is that both the page title and the filename are just bait. The attackers simply use popular search terms to lure users into downloading the malware. The actual file content has nothing to do with what’s advertised — inside, it’s always the same infostealer.

The site also claimed that all hosted files were scanned for viruses. When the user decides to download, say, a pirated game, the site displays a banner saying the file is being scanned by various antivirus engines. Of course, no such scanning actually takes place; the attackers are merely trying to create an illusion of trustworthiness.

What makes Stealka dangerous

Stealka has a fairly extensive arsenal of capabilities, but its prime target is data from browsers built on the Chromium and Gecko engines. This puts over a hundred different browsers at risk, including popular ones like Chrome, Firefox, Opera, Yandex Browser, Edge, Brave, as well as many, many others.

Browsers store a huge amount of sensitive information, which attackers use to hijack accounts and continue their attacks. The main targets are autofill data, such as sign-in credentials, addresses, and payment card details. We’ve warned repeatedly that saving passwords in your browser is risky — attackers can extract them in seconds. Cookies and session tokens are perhaps even more valuable to hackers, as they can allow criminals to bypass two-factor authentication and hijack accounts without entering the password.

The story doesn’t end with the account hack. Attackers use these compromised accounts to spread the malware further. For example, we discovered the stealer in a GTAV mod posted on a dedicated site by an account that had previously been compromised.

Beyond stealing browser data, Stealka also targets the settings and databases of 115 browser extensions for crypto wallets, password managers, and 2FA services. Here are some of the most popular extensions now at risk:

• Crypto wallets: Binance, Coinbase, Crypto.com, SafePal, Trust Wallet, MetaMask, Ton, Phantom, Exodus
• Two-factor authentication: Authy, Google Authenticator, Bitwarden
• Password management: 1Password, Bitwarden, LastPass, KeePassXC, NordPass

Finally, the stealer also downloads local settings, account data, and service files from a wide variety of applications:

• Crypto wallets. Wallet configurations may contain encrypted private keys, seed-phrase data, wallet file paths, and encryption parameters. That’s enough to at least make an attempt at stealing your cryptocurrency. At risk are 80 wallet applications, including Binance, Bitcoin, BitcoinABC, Dogecoin, Ethereum, Exodus, Mincoin, MyCrypto, MyMonero, Monero, Nexus, Novacoin, Solar, and many others.
• Messaging apps. Messaging app service files store account data, device identifiers, authentication tokens, and the encryption parameters for your conversations. In theory, a malicious actor could gain access to your account and read your chats. At risk are Discord, Telegram, Unigram, Pidgin, Tox, and others.
• Password managers. Even if the passwords themselves are encrypted, the configuration files often contain information that makes cracking the vault significantly easier: encryption parameters, synchronization tokens, and details about the vault version and structure. At risk are 1Password, Authy, Bitwarden, KeePass, LastPass, and NordPass.
• Email clients. These are where your account credentials, mail server connection settings, authentication tokens, and local copies of your emails can be found. With access to your email, an attacker will almost certainly attempt to reset passwords for your other services. At risk are Gmail Notifier Pro, Claws, Mailbird, Outlook, Postbox, The Bat!, Thunderbird, and TrulyMail.
• Note-taking apps. Instead of shopping lists or late-night poetry, some users store information in their notes that has no business being there, like seed phrases or passwords. At risk are NoteFly, Notezilla, SimpleStickyNotes, and Microsoft StickyNotes.
• Gaming services and clients. The local files of gaming platforms and launchers store account data, linked service information, and authentication tokens. At risk are Steam, Roblox, Intent Launcher, Lunar Client, TLauncher, Feather Client, Meteor Client, Impact Client, Badlion Client, and WinAuth for battle.net.
• VPN clients. By gaining access to configuration files, attackers can hijack the victim’s VPN account to mask their own malicious activities. At risk are AzireVPN, OpenVPN, ProtonVPN, Surfshark, and WindscribeVPN.

That’s an extensive list — and we haven’t even named all of them! In addition to local files, this infostealer also harvests general system data: a list of installed programs, the OS version and language, username, computer hardware information, and miscellaneous settings. And as if that weren’t enough, the malware also takes screenshots.

How to protect yourself from Stealka and other infostealers

• Secure your device with reliable antivirus software. Even downloading files from legitimate websites is no guarantee of safety — attackers leverage trusted platforms to distribute stealers all the time. Kaspersky Premium detects malware on your computer in time and alerts you to the threat.
• Don’t store sensitive information in browsers. It’s handy — no one can argue with that. But unfortunately browsers aren’t the most secure environment for your data. Sign-in credentials, bank card details, secret notes, and other confidential information are better kept in a securely encrypted format in Kaspersky Password Manager, which is immune to the exploits used by Stealka.
• Be careful with game cheats, mods, and especially pirated software. It’s better to pay up for official software than to chase the false savings offered by software cracks, and end up losing all your money.
• Enable two-factor authentication or use backup codes wherever possible. Two-factor authentication (2FA) makes life much harder for attackers, while backup codes help you regain access to your critical accounts if compromised. Just be sure not to store backup codes in text documents, notes, or your browser. For all your backup codes and 2FA tokens, use a reliable password manager.

Curious what other stealers are out there, and what they’re capable of? Read more in our other posts:

• Beware of stealers disguised as… wedding invitations
• AMOS infostealer distributed via ChatGPT chat-sharing feature
• Banshee: A stealer targeting macOS users
• Arcane stealer instead of Minecraft cheats
• Efimer Trojan using hacked websites to steal cryptocurrency

Admit it: you’ve been meaning to jump on the latest NFT reincarnation — Telegram Gifts — but just haven’t gotten around to it. It’s the hottest trend right now. Developers are churning out collectible images in partnership with celebs like Snoop Dogg. All your friends’ profiles are already decked out with these modish pictures, and you’re dying to hop on this hype train — but pay as little as possible for it.

And then it happens — a stranger messages you privately with a generous offer: a chance to snag a couple of these digital gifts — with no investment required. A bot that looks completely legit is running an airdrop. In the world of NFTs, an airdrop is a promotional stunt where a small number of new crypto assets are given away for free. The buzzword has been adopted on Telegram, thanks to the crypto nature of these gifts and the NFT mechanics running under the hood.

They’re offering you these gift images for free — or so they say. You could later attach them to your profile or sell them for Telegram’s native currency, Toncoin. You don’t even have to tap an external link. Just hit a button in the message, launch a Mini App right inside Telegram itself, and enter your login credentials. And then… your account immediately gets hijacked. You won’t get any gifts, and overall, you’ll be left with anything but a celebratory feeling.

Today, we break down a phishing scheme that exploits Telegram’s built-in Mini Apps, and share tips to help you avoid falling for these attacks.

How the new phishing scheme works

The principle of classic phishing is straightforward: the user gets a link to a fake website that mimics a legitimate sign-in form. When the victim enters their credentials, this data goes straight to the scammer. However, phishing tactics are constantly evolving, and this new attack method is far more insidious.

The bad actors create phishing Mini Apps directly inside Telegram. These appear as standard web pages but are embedded within the messaging app’s interface instead of opening in an external browser. To the user, these apps look completely legitimate. After all, they run within the official Telegram app itself.

This leads the victim to think, “If this app runs inside Telegram, there must be some kind of vetting process for these apps. Surely they wouldn’t let an obvious scam through?” In practice, it turns out that’s not the case at all.

How is this scheme even a thing?

A core security issue with Telegram Mini Apps is that the platform does almost no vetting before an app goes live. This is a world apart from the strict review processes used by Google Play and the App Store — although even there, obvious malware occasionally slips through.

On Telegram, it’s far easier for bad actors. Essentially, anyone who wishes to create and launch a Mini App can do so. Telegram does not review the code, functionality, or the developer’s intent. This turns a security flaw within a messaging service boasting nearly a billion global users into a global-scale problem. To make matters worse, moderation of these Mini Apps within Telegram is entirely reactive — meaning action is only taken after users start complaining or law enforcement gets involved.

In this case, the bait was “gifts” from UFC fighters: a giveaway of “papakhas” — digital gift images of the traditional Dagestani hat released by Telegram in partnership with Khabib Nurmagomedov. An auction for these items did take place, with Pavel Durov even posting about it on his X and Telegram (Khabib reposted these announcements but later deleted them after the auction ended). However, there were only 29 000 of these “papakhas” released, which wasn’t enough to satisfy all the eager fans. Scammers seized on the opportunity, assuring fans they could get the exclusive items for free. The phishing campaign was a targeted one — focusing on users who’d been active on the athlete’s channel.

How the scammers lull their victims

The criminals leveraged the name of the popular Portals platform — a legitimate service for games, apps, and entertainment within Telegram. They created a series of Mini Apps that were visually almost indistinguishable from the real ones, and promoted them as free giveaways — airdrops.

That said, the scam campaigns themselves show signs of being rushed and cutting design and copywriting costs — with obvious signs of AI involvement. Some of the messages contain leftover text fragments clearly generated by a neural network, which the scammers either forgot or couldn’t be bothered to edit.

How to protect your Telegram account from being hacked

The golden security rules are simple: stay vigilant, and learn the key hallmarks of these attacks:

• Verify the source. If you receive a link promising a giveaway from a celebrity or even Telegram itself but sent from an unfamiliar account or a dubious group, don’t click. Cross-check through the celebrity or company’s official channel to see if they’re actually running a promo like that.
• Inspect the account verification badge. Ascertain that the blue checkmark is real and not just an emoji status or part of the profile name. You can verify this by simply tapping that checkmark icon in the profile. If it’s a Premium emoji status, Telegram will explicitly tell you so. If a checkmark emoji is simply added to the profile name, tapping it doesn’t do anything. But if the account is genuinely verified, tapping the blue checkmark will bring up an official confirmation message from Telegram.
• Don’t be in a rush to authenticate in Mini Apps. Legitimate Telegram apps typically don’t require you to sign in again through a form inside the Mini App. If you’re prompted to enter your phone number or a verification code, it’s likely a phishing attempt.
• Look for signs of AI-generated text or design. Weird grammar, unnatural phrasing, or leftover neural network prompts within a message are a red flag. Scammers frequently use AI-powered generation to churn out text quickly and cheaply.
• Turn on two-step verification (your Telegram password). Do this right now in Settings → Privacy and Security → Two-Step Verification. Even if a scammer manages to get your phone number and SMS code, they won’t be able to access your account without this password. Obviously, never share your password with anyone — it’s meant only for you to sign in to your Telegram account.
• Use a passkey to secure your account. A recent Telegram update added the ability to securely sign in with a passkey. We’ve covered using passkeys with popular services and the associated caveats in detail. A passkey makes it nearly impossible for a malicious actor to steal your account. You can set one up in Settings → Privacy and Security → Passkeys.
• Store your password and passkey in a password manager. If you’ve secured your account with both a password and a passkey, remember that a weak, reused, or compromised password can still be the proverbial “spare key under the mat” for attackers — even if the “front door” is locked with a passkey. Therefore, we recommend creating a strong, unique password for Telegram and storing it — along with your passkey — in Kaspersky Password Manager. This keeps your credentials and keys available across all your devices.
• Install Kaspersky for Android on your smartphone. Its new anti-phishing technology protects you from phishing links embedded in notifications from any app.

What to do if your Telegram account was already stolen

The key is keeping calm and acting swiftly. You have just 24 hours to reclaim your account, or you risk losing it permanently. Follow the step-by-step guide to restoring access in our post What to do if your Telegram account is hacked.

Finally, a reminder that has become our classic mantra: if an offer looks too good to be true, it almost certainly is. Always verify information through official channels, and never enter your passwords or passkeys into unofficial apps or forms — even if they look legit. Stay vigilant and stay safe.

Want more tips on securing your messenger accounts and chats? Check out our related posts:

• Messengers 101: safety and privacy advice
• Messaging other platforms via WhatsApp: the pros and cons
• WhatsApp and Telegram account hijacking: How to protect yourself against scams
• What to do if your WhatsApp account gets hacked
• What to do if your Telegram account is hacked

In August 2025, we discovered a campaign targeting individuals in Turkey with a new Android banking Trojan we dubbed “Frogblight”. Initially, the malware was disguised as an app for accessing court case files via an official government webpage. Later, more universal disguises appeared, such as the Chrome browser.

Frogblight can use official government websites as an intermediary step to steal banking credentials. Moreover, it has spyware functionality, such as capabilities to collect SMS messages, a list of installed apps on the device and device filesystem information. It can also send arbitrary SMS messages.

Another interesting characteristic of Frogblight is that we’ve seen it updated with new features throughout September. This may indicate that a feature-rich malware app for Android is being developed, which might be distributed under the MaaS model.

This threat is detected by Kaspersky products as HEUR:Trojan-Banker.AndroidOS.Frogblight.*, HEUR:Trojan-Banker.AndroidOS.Agent.eq, HEUR:Trojan-Banker.AndroidOS.Agent.ep, HEUR:Trojan-Spy.AndroidOS.SmsThief.de.

Technical details

Background

While performing an analysis of mobile malware we receive from various sources, we discovered several samples belonging to a new malware family. Although these samples appeared to be still under development, they already contained a lot of functionality that allowed this family to be classified as a banking Trojan. As new versions of this malware continued to appear, we began monitoring its development. Moreover, we managed to discover its control panel and based on the “fr0g” name shown there, we dubbed this family “Frogblight”.

Initial infection

We believe that smishing is one of the distribution vectors for Frogblight, and that the users had to install the malware themselves. On the internet, we found complaints from Turkish users about phishing SMS messages convincing users that they were involved in a court case and containing links to download malware. versions of Frogblight, including the very first ones, were disguised as an app for accessing court case files via an official government webpage and were named the same as the files for downloading from the links mentioned above.

While looking for online mentions of the names used by the malware, we discovered one of the phishing websites distributing Frogblight, which disguises itself as a website for viewing a court file.

We were able to open the admin panel of this website, where it was possible to view statistics on Frogblight malware downloads. However, the counter had not been fully implemented and the threat actor could only view the statistics for their own downloads.

Additionally, we found the source code of this phishing website available in a public GitHub repository. Judging by its description, it is adapted for fast deployment to Vercel, a platform for hosting web apps.

App features

As already mentioned, Frogblight was initially disguised as an app for accessing court case files via an official government webpage. Let’s look at one of the samples using this disguise (9dac23203c12abd60d03e3d26d372253). For analysis, we selected an early sample, but not the first one discovered, in order to demonstrate more complete Frogblight functionality.

After starting, the app prompts the victim to grant permissions to send and read SMS messages, and to read from and write to the device’s storage, allegedly needed to show a court file related to the user.

The full list of declared permissions in the app manifest file is shown below:

• MANAGE_EXTERNAL_STORAGE
• READ_EXTERNAL_STORAGE
• WRITE_EXTERNAL_STORAGE
• READ_SMS
• RECEIVE_SMS
• SEND_SMS
• WRITE_SMS
• RECEIVE_BOOT_COMPLETED
• INTERNET
• QUERY_ALL_PACKAGES
• BIND_ACCESSIBILITY_SERVICE
• DISABLE_KEYGUARD
• FOREGROUND_SERVICE
• FOREGROUND_SERVICE_DATA_SYNC
• POST_NOTIFICATIONS
• QUICKBOOT_POWERON
• RECEIVE_MMS
• RECEIVE_WAP_PUSH
• REQUEST_IGNORE_BATTERY_OPTIMIZATIONS
• SCHEDULE_EXACT_ALARM
• USE_EXACT_ALARM
• VIBRATE
• WAKE_LOCK
• ACCESS_NETWORK_STATE
• READ_PHONE_STATE

After all required permissions are granted, the malware opens the official government webpage for accessing court case files in WebView, prompting the victim to sign in. There are different sign-in options, one of them via online banking. If the user chooses this method, they are prompted to click on a bank whose online banking app they use and fill out the sign-in form on the bank’s official website. This is what Frogblight is after, so it waits two seconds, then opens the online banking sign-in method regardless of the user’s choice. For each webpage that has finished loading in WebView, Frogblight injects JavaScript code allowing it to capture user input and send it to the C2 via a REST API.

The malware also changes its label to “Davalarım” if the Android version is newer than 12; otherwise it hides the icon.

The app icon before (left) and after launching (right)
In the sample we review in this section, Frogblight uses a REST API for C2 communication, implemented using the Retrofit library. The malicious app pings the C2 server every two seconds in foreground, and if no error is returned, it calls the REST API client methods fetchOutbox and getFileCommands. Other methods are called when specific events occur, for example, after the device screen is turned on, the com.capcuttup.refresh.PersistentService foreground service is launched, or an SMS is received. The full list of all REST API client methods with parameters and descriptions is shown below.

REST API client method	Description	Parameters
fetchOutbox	Request message content to be sent via SMS or displayed in a notification	device_id: unique Android device ID
ackOutbox	Send the results of processing a message received after calling the API method fetchOutbox	device_id: unique Android device ID msg_id: message ID status: message processing status error: message processing error
getAllPackages	Request the names of app packages whose launch should open a website in WebView to capture user input data	action: same as the API method name
getPackageUrl	Request the website URL that will be opened in WebView when the app with the specified package name is launched	action: same as the API method name package: the package name of the target app
getFileCommands	Request commands for file operations Available commands: ●       download: upload the target file to the C2 ●       generate_thumbnails: generate thumbnails from the image files in the target directory and upload them to the C2 ●       list: send information about all files in the target directory to the C2 ●       thumbnail: generate a thumbnail from the target image file and upload it to the C2	device_id: unique Android device ID
pingDevice	Check the C2 connection	device_id: unique Android device ID
reportHijackSuccess	Send captured user input data from the website opened in a WebView when the app with the specified package name is launched	action: same as the API method name package: the package name of the target app data: captured user input data
saveAppList	Send information about the apps installed on the device	device_id: unique Android device ID app_list: a list of apps installed on the device app_count: a count of apps installed on the device
saveInjection	Send captured user input data from the website opened in a WebView. If it was not opened following the launch of the target app, the app_name parameter is determined based on the opened URL	device_id: unique Android device ID app_name: the package name of the target app form_data: captured user input data
savePermission	Unused but presumably needed for sending information about permissions	device_id: unique Android device ID permission_type: permission type status: permission status
sendSms	Send information about an SMS message from the device	device_id: unique Android device ID sender: the sender’s/recipient’s phone number message: message text timestamp: received/sent time type: message type (inbox/sent)
sendTelegramMessage	Send captured user input data from the webpages opened by Frogblight in WebView	device_id: unique Android device ID url: website URL title: website page title input_type: the type of user input data input_value: user input data final_value: user input data with additional information timestamp: the time of data capture ip_address: user IP address sms_permission: whether SMS permission is granted file_manager_permission: whether file access permission is granted
updateDevice	Send information about the device	device_id: unique Android device ID model: device manufacturer and model android_version: Android version phone_number: user phone number battery: current battery level charging: device charging status screen_status: screen on/off ip_address: user IP address sms_permission: whether SMS permission is granted file_manager_permission: whether file access permission is granted
updatePermissionStatus	Send information about permissions	device_id: unique Android device ID permission_type: permission type status: permission status timestamp: current time
uploadBatchThumbnails	Upload thumbnails to the C2	device_id: unique Android device ID thumbnails: thumbnails
uploadFile	Upload a file to the C2	device_id: unique Android device ID file_path: file path download_id: the file ID on the C2 The file itself is sent as an unnamed parameter
uploadFileList	Send information about all files in the target directory	device_id: unique Android device ID path: directory path file_list: information about the files in the target directory
uploadFileListLog	Send information about all files in the target directory to an endpoint different from uploadFileList	device_id: unique Android device ID path: directory path file_list: information about the files in the target directory
uploadThumbnailLog	Unused but presumably needed for uploading thumbnails to an endpoint different from uploadBatchThumbnails	device_id: unique Android device ID thumbnails: thumbnails

Remote device control, persistence, and protection against deletion

The app includes several classes to provide the threat actor with remote access to the infected device, gain persistence, and protect the malicious app from being deleted.

• capcuttup.refresh.AccessibilityAutoClickService
  This is intended to prevent removal of the app and to open websites specified by the threat actor in WebView upon target apps startup. It is present in the sample we review, but is no longer in use and deleted in further versions.
• capcuttup.refresh.PersistentService
  This is a service whose main purpose is to interact with the C2 and to make malicious tasks persistent.
• capcuttup.refresh.BootReceiver
  This is a broadcast receiver responsible for setting up the persistence mechanisms, such as job scheduling and setting alarms, after device boot completion.

Further development

In later versions, new functionality was added, and some of the more recent Frogblight variants disguised themselves as the Chrome browser. Let’s look at one of the fake Chrome samples (d7d15e02a9cd94c8ab00c043aef55aff).

In this sample, new REST API client methods have been added for interacting with the C2.

REST API client method	Description	Parameters
getContactCommands	Get commands to perform actions with contacts Available commands: ●       ADD_CONTACT: add a contact to the user device ●       DELETE_CONTACT: delete a contact from the user device ●       EDIT_CONTACT: edit a contact on the user device	device_id: unique Android device ID
sendCallLogs	Send call logs to the C2	device_id: unique Android device ID call_logs: call log data
sendNotificationLogs	Send notifications log to the C2. Not fully implemented in this sample, and as of the time of writing this report, we hadn’t seen any samples with a full-fledged implementation of this API method	action: same as the API method name notifications: notification log data

Also, the threat actor had implemented a custom input method for recording keystrokes to a file using the com.puzzlesnap.quickgame.CustomKeyboardService service.

Another Frogblight sample we observed trying to avoid emulators and using geofencing techniques is 115fbdc312edd4696d6330a62c181f35. In this sample, Frogblight checks the environment (for example, device model) and shuts down if it detects an emulator or if the device is located in the United States.

Later on, the threat actor decided to start using a web socket instead of the REST API. Let’s see an example of this in one of the recent samples (08a3b1fb2d1abbdbdd60feb8411a12c7). This sample is disguised as an app for receiving social support via an official government webpage. The feature set of this sample is very similar to the previous ones, with several new capabilities added. Commands are transmitted over a web socket using the JSON format. A command template is shown below:

{
    "id": <command ID>,
    "command_type": <command name>
    "command_data": <command data>
}

It is also worth noting that some commands in this version share the same meaning but have different structures, and the functionality of certain commands has not been fully implemented yet. This indicates that Frogblight was under active development at the time of our research, and since no its activity was noticed after September, it is possible that the malware is being finalized to a fully operational state before continuing to infect users’ devices. A full list of commands with their parameters and description is shown below:

Command	Description	Parameters
connect	Send a registration message to the C2	–
connection_success	Send various information, such as call logs, to the C2; start pinging the C2 and requesting commands	–
auth_error	Log info about an invalid login key to the Android log system	–
pong_device	Does nothing	–
commands_list	Execute commands	List of commands
sms_send_command	Send an arbitrary SMS message	recipient: message destination message: message text msg_id: message ID
bulk_sms_command	Send an arbitrary SMS message to multiple recipients	recipients: message destinations message: message text
get_contacts_command	Send all contacts to the C2	–
get_app_list_command	Send information about the apps installed on the device to the C2	–
get_files_command	Send information about all files in certain directories to the C2	–
get_call_logs_command	Send call logs to the C2	–
get_notifications_command	Send a notifications log to the C2. This is not fully implemented in the sample at hand, and as of the time of writing this report, we hadn’t seen any samples with a full-fledged implementation of this command	–
take_screenshot_command	Take a screenshot. This is not fully implemented in the sample at hand, and as of the time of writing this report, we hadn’t seen any samples with a full-fledged implementation of this command	–
update_device	Send registration message to the C2	–
new_webview_data	Collect WebView data. This is not fully implemented in the sample at hand, and as of the time of writing this report, we hadn’t seen any samples with a full-fledged implementation of this command	–
new_injection	Inject code. This is not fully implemented in the sample at hand, and as of the time of writing this report, we hadn’t seen any samples with a full-fledged implementation of this command	code: injected code target_app: presumably the package name of the target app
add_contact_command	Add a contact to the user device	name: contact name phone: contact phone email: contact email
contact_add	Add a contact to the user device	display_name: contact name phone_number: contact phone email: contact email
contact_delete	Delete a contact from the user device	phone_number: contact phone
contact_edit	Edit a contact on the user device	display_name: new contact name phone_number: contact phone email: new contact email
contact_list	Send all contacts to the C2	–
file_list	Send information about all files in the specified directory to the C2	path: directory path
file_download	Upload the specified file to the C2	file_path: file path download_id: an ID that is received with the command and sent back to the C2 along with the requested file. Most likely, this is used to organize data on the C2
file_thumbnail	Generate a thumbnail from the target image file and upload it to the C2	file_path: image file path
file_thumbnails	Generate thumbnails from the image files in the target directory and upload them to the C2	folder_path: directory path
health_check	Send information about the current device state: battery level, screen state, and so on	–
message_list_request	Send all SMS messages to the C2	–
notification_send	Show an arbitrary notification	title: notification title message: notification message app_name: notification subtext
package_list_response	Save the target package names	packages: a list of all target package names. Each list element contains: package_name: target package name active: whether targeting is active
delete_contact_command	Delete a contact from the user device. This is not fully implemented in the sample at hand, and as of the time of writing this report, we hadn’t seen any samples with a full-fledged implementation of this command	contact_id: contact ID name: contact name
file_upload_command	Upload specified file to the C2. This is not fully implemented in the sample at hand, and as of the time of writing this report, we hadn’t seen any samples with a full-fledged implementation of this command	file_path: file path file_name: file name
file_download_command	Download file to user device. This is not fully implemented in the sample at hand, and as of the time of writing this report, we hadn’t seen any samples with a full-fledged implementation of this command	file_url: the URL of the file to download download_path: download path
download_file_command	Download file to user device. This is not fully implemented in the sample at hand, and as of the time of writing this report, we hadn’t seen any samples with a full-fledged implementation of this command	file_url: the URL of the file to download download_path: downloading path
get_permissions_command	Send a registration message to the C2, including info about specific permissions	–
health_check_command	Send information about the current device state, such as battery level, screen state, and so on	–
connect_error	Log info about connection errors to the Android log system	A list of errors
reconnect	Send a registration message to the C2	–
disconnect	Stop pinging the C2 and requesting commands from it	–

Authentication via WebSocket takes place using a special key.

At the IP address to which the WebSocket connection was made, the Frogblight web panel was accessible, which accepted the authentication key mentioned above. Since only samples using the same key as the webpanel login are controllable through it, we suggest that Frogblight might be distributed under the MaaS model.

Judging by the menu options, the threat actor can sort victims’ devices by certain parameters, such as the presence of banking apps on the device, and send bulk SMS messages and perform other mass actions.

Victims

Since some versions of Frogblight opened the Turkish government webpage to collect user-entered data on Turkish banks’ websites, we assume with high confidence that it is aimed mainly at users from Turkey. Also, based on our telemetry, the majority of users attacked by Frogblight are located in that country.

Attribution

Even though it is not possible to provide an attribution to any known threat actor based on the information available, during our analysis of the Frogblight Android malware and the search for online mentions of the names it uses, we discovered a GitHub profile containing repos with Frogblight, which had also created repos with Coper malware, distributed under the MaaS model. It is possible that this profile belongs to the attackers distributing Coper who have also started distributing Frogblight.

Also, since the comments in the Frogblight code are written in Turkish, we believe that its developers speak this language.

Conclusions

The new Android malware we dubbed “Frogblight” appeared recently and targets mainly users from Turkey. This is an advanced banking Trojan aimed at stealing money. It has already infected real users’ devices, and it doesn’t stop there, adding more and more new features in the new versions that appear. It can be made more dangerous by the fact that it may be used by attackers who already have experience distributing malware. We will continue to monitor its development.

Indicators of Compromise

More indicators of compromise, as well as any updates to these, are available to the customers of our crimeware reporting service. If you are interested, please contact crimewareintel@kaspersky.com.

APK file hashes
8483037dcbf14ad8197e7b23b04aea34
105fa36e6f97977587a8298abc31282a
e1cd59ae3995309627b6ab3ae8071e80
115fbdc312edd4696d6330a62c181f35
08a3b1fb2d1abbdbdd60feb8411a12c7
d7d15e02a9cd94c8ab00c043aef55aff
9dac23203c12abd60d03e3d26d372253

C2 domains
1249124fr1241og5121.sa[.]com
froglive[.]net

C2 IPs
45.138.16.208[:]8080

URL of GitHub repository with Frogblight phishing website source code
https://github[.]com/eraykarakaya0020/e-ifade-vercel

URL of GitHub account containing APK files of Frogblight and Coper
https://github[.]com/Chromeapk

Distribution URLs
https://farketmez37[.]cfd/e-ifade.apk
https://farketmez36[.]sbs/e-ifade.apk
https://e-ifade-app-5gheb8jc.devinapps[.]com/e-ifade.apk

We discuss the CVSS 10.0-rated RCE vulnerability in the Flight protocol used by React Server Components. This is tracked as CVE-2025-55182.

The post Exploitation of Critical Vulnerability in React Server Components (Updated December 12) appeared first on Unit 42.

Self-replicating worm “Shai-Hulud” has compromised hundreds of software packages in a supply chain attack targeting the npm ecosystem. We discuss scope and more.

The post "Shai-Hulud" Worm Compromises npm Ecosystem in Supply Chain Attack (Updated November 26) appeared first on Unit 42.