Fake CAPTCHA pages keep scanners out and get victims to run the malware themselves
Barracuda found CAPTCHA abuse in 43% of the phishing-kit attacks it examined during 2025. The check can keep automated scanners from reaching the next page, so a URL scan may inspect only the CAPTCHA.
Reality
- Evidence45
- Adoption43
- Hype gap+10
- Incentives40
- Confidence55
CERT-UA says attackers planted fake Cloudflare checks on more than 100 legitimate websites to get visitors to install Lunex Stealer themselves. Ontinue says its browser components can keep file access after the stealer is deleted, so cleanup has to reach the browsers.
Perspective Coverage
3 publishers
- Builder
- Builder 30%
- Operator
- Operator 62%
- Investor
- Investor 8%
Reality
- Evidence72
- Adoption
- Not scored
- Hype gap+10
- Incentives25
- Confidence70
Microsoft documented a ClickFix variant that stages its payload in the browser cache, keeping the command the victim pastes under Windows' 260-character Run limit. Detection moves to script engines reading browser profile folders, not the pasted text.
Reality
- Evidence70
- Adoption
- Not scored
- Hype gap+5
- Incentives30
- Confidence72
Huntress traced a RAT campaign that hides its first-stage lure in a ChatGPT Custom GPT on the real chatgpt.com, across at least 40 incidents. Because the page sits on a trusted domain, blocking the ClickFix PowerShell paste is the control that works.
Perspective Coverage
8 publishers
- Builder
- Builder 30%
- Operator
- Operator 64%
- Investor
- Investor 6%
Reality
- Evidence70
- Adoption20
- Hype gap+30
- Incentives40
- Confidence68
Fake custom GPTs on chatgpt.com open an eight-stage ClickFix malware chain
Huntress traced at least two of more than 40 ClickFix malware incidents to fake custom GPTs on OpenAI's chatgpt.com. On that route every hop before the PowerShell command sat on a Google or OpenAI address, so staff taught to trust familiar domains would click through each one.
Reality
- Evidence45
- Adoption20
- Hype gap+20
- Incentives
- Not scored
- Confidence50
Lunex sells a stealer that blinds security tools with a vulnerable AMD driver
Ontinue says the Lunex stealer-for-rent loads a flawed AMD driver, CVE-2023-20598, to switch off security tools before robbing seven browsers. That driver step is rarely used ahead of an info stealer, and Lunex is rented to multiple criminal groups.
Reality
- Evidence55
- Adoption40
- Hype gap+10
- Incentives
- Not scored
- Confidence55
Jamf Threat Labs says the Rust macOS stealer pastes itself in via Terminal, fakes an Installer password prompt, mutes the speakers, then hands the operator a headless clone of your logged-in profile.
Perspective Coverage
5 publishers
- Builder
- Builder 30%
- Operator
- Operator 65%
- Investor
- Investor 5%
Reality
- Evidence68
- Adoption
- Not scored
- Hype gap+10
- Incentives40
- Confidence70
One Polygon smart contract rotated ClickFix lure hosts three times in a single day
CTM360 tracked more than 17,000 compromised URLs serving a fake Cloudflare check whose next hostname arrives from an on-chain lookup. Microsoft put 47 percent of its 2025 initial-access cases down to the technique.
Reality
- Evidence45
- Adoption72
- Hype gap+20
- Incentives70
- Confidence50
Cato Networks says a sponsored ad outranked OpenAI's own page, sending macOS developers to a Google Sites portal whose ClickFix content lives on infrastructure Google never hosted.
Perspective Coverage
3 publishers
- Builder
- Builder 34%
- Operator
- Operator 58%
- Investor
- Investor 8%
Reality
- Evidence60
- Adoption
- Not scored
- Hype gap+10
- Incentives35
- Confidence65
OX Security says the packages were never meant to infect anyone who installs them. Mirrored through unpkg, they serve attacker HTML from a domain most egress policies wave through.
Reality
- Evidence62
- Adoption
- Not scored
- Hype gap+20
- Incentives
- Not scored
- Confidence65
Hudson Rock and ADAMnetworks link the ads to an operation they call PasteSwitch, which hands Windows and macOS visitors different paste-this-command lures and swaps payloads on the backend depending on who arrives.
Perspective Coverage
5 publishers
- Builder
- Builder 29%
- Operator
- Operator 65%
- Investor
- Investor 6%
Reality
- Evidence74
- Adoption
- Not scored
- Hype gap+8
- Incentives50
- Confidence72
Arctic Wolf Labs found injected iframes on real Ukrainian business sites showing a Ukrainian-language fake Cloudflare check that copies a Windows Installer command for the visitor to paste into Run. Its exposed panel logged 557 views.
Perspective Coverage
3 publishers
- Builder
- Builder 30%
- Operator
- Operator 64%
- Investor
- Investor 6%
Reality
- Evidence68
- Adoption14
- Hype gap+8
- Incentives35
- Confidence64
Blackpoint says a fake Spotify MSI, launched through msiexec.exe, ends with a Node.js agent that reads its live C2 address off a Polygon contract. The same implant has also shipped as Zoom and Teams.
Reality
- Evidence62
- Adoption
- Not scored
- Hype gap+10
- Incentives
- Not scored
- Confidence66
Kaspersky says a new MacSync variant hides its payload commands in public iCloud calendar events and downloads the next stage from iCloud. Both the instructions and the download ride trusted Apple domains, past defenses that block known-bad hosts.
Reality
- Evidence70
- Adoption
- Not scored
- Hype gap0
- Incentives
- Not scored
- Confidence68
Exvicy built its ClickFix subscription on code lifted from rival ErrTraffic
Sekoia's comparison found the two ClickFix services' injected scripts near identical, while the Exvicy subscription climbed from $1,200 to $2,000 a month and its panel count reached about 80 hosts by late August.
Reality
- Evidence62
- Adoption35
- Hype gap+10
- Incentives40
- Confidence58
ClickFix crews are staging the second stage inside Binance Smart Chain smart contracts
Blackpoint's SOC worked an incident where a fake captcha on the victim company's own WordPress site ran a hidden PowerShell downloader, and the campaign's later stage sits in a blockchain contract no provider can pull.
Publishers:blackpointcyber.com
Reality
- Evidence60
- Adoption30
- Hype gap+15
- Incentives65
- Confidence55
FakeAgent delivered SectopRAT to more than 29 organisations from a page hosted on claude.ai
Huntress tracked nine months of malware campaigns run through shipped AI sharing features. Public Claude Artifacts, claude.ai/share links and indexable ChatGPT and Grok conversations all put the lure on a domain the victim already trusts.
Reality
- Evidence45
- Adoption38
- Hype gap+12
- Incentives85
- Confidence50
SloppyRAT parks its backup C2 channel on the Polygon blockchain
Zscaler's ThreatLabz found SloppyRAT in June 2026 at the end of a ClickFix chain whose live hosts are all ordinary domains, with a Polygon JSON-RPC lookup held in reserve for when those domains stop answering.
Reality
- Evidence62
- Adoption18
- Hype gap+18
- Incentives60
- Confidence55
Talos ties a fake Google CAPTCHA to a DLL executing from a remote WebDAV share
Cisco Talos found the loader running from a WebDAV UNC path at a Ukrainian government organisation in April 2026, delivered by a ClickFix prompt whose JavaScript was stored on BNB Smart Chain. The primary payload is Amatera stealer.
Reality
- Evidence68
- Adoption38
- Hype gap−12
- Incentives55
- Confidence60
5,400 hacked WordPress and PrestaShop sites pull their ClickFix payload from a BNB testnet contract
The contract cannot be deleted. Every injected site has to reach a public BNB Smart Chain testnet RPC endpoint to read it. Netskope recommends blocking that pool. The payload has already changed once.
Reality
- Evidence52
- Adoption64
- Hype gap+14
- Incentives58
- Confidence55