ClickFix Campaign Uses AppleScript Stealer to Target macOS Credentials and Session Cookies

· Original article ↗

Summary

Netskope details a ClickFix campaign that tricks macOS users into running a command, then uses a persistent fake password prompt to steal credentials, keychain and browser data, extension contents, and cryptocurrency wallet files.

Key points

  • A landing page filters out mobile users and routes desktop victims to Windows or macOS payloads based on their user-agent.
  • A fake CAPTCHA instructs victims to paste and run a command that downloads and silently executes the macOS script.
  • The AppleScript displays a persistent, convincing password dialog and validates entered credentials in real time.
  • The stealer targets the macOS Keychain, browser cookies and saved data, data from more than 200 extensions, and files from 16 cryptocurrency wallets.
  • Stolen data is compressed into a ZIP archive and sent to the attacker’s server over HTTP; the script then removes its staging files.
  • Stolen session cookies can enable account access by hijacking active sessions and bypassing MFA.
  • Netskope recommends updating to macOS Tahoe 26.4 or macOS Sequoia, which adds a Terminal warning about pasting potentially malicious commands.

Article Details

Attack Vectors
  • A fake CAPTCHA page uses client-side JavaScript to filter for desktop users and direct macOS users to an AppleScript-based loader.
  • Victims are tricked into copying and pasting a malicious command into Terminal, which downloads and runs the script in the background.
  • A persistent deceptive AppleScript dialog prompts victims for their system password and repeatedly reappears until a valid password is entered.
  • The script collects Keychain data, browser credentials and session cookies, extension data, and files from standalone cryptocurrency wallets, then exfiltrates the collected data.
Defensive Notes
  • The article recommends updating to macOS Tahoe 26.4 or macOS Sequoia, whose Terminal security warning alerts users before they paste potentially harmful commands from an untrusted source.
  • Treat commands presented by fake CAPTCHA pages or other untrusted sources as suspicious; the described attack depends on victims manually pasting and executing the command.

Indicators of compromise

TypeIndicatorContext
URLhxxp[:]//172[.]94[.]9[.]250Attacker's command-and-control server; the article says stolen data is exfiltrated to it via HTTP POST.

MITRE ATT&CK

Products

Related Articles