Threat reportMalwareTL-2026-2441
LegionLoader Malware Distributed via Fake Cloudflare CAPTCHA Using the ClickFix Technique
AhnLab ASEC identified an active campaign distributing LegionLoader, a multi-stage loader/backdoor, via the ClickFix social-engineering technique. Victims reach a fake Cloudflare CAPTCHA page — through a compromised South Korean namu.wiki page or spear-phishing emails impersonating internal account-provisioning systems — that tricks them into pasting and executing malicious PowerShell into the Windows Run dialog or a terminal.
Key facts
| Identifiers | TL-2026-2441 |
|---|---|
| Also known as | Legion Loader |
| Severity | MEDIUM |
| Affected products | Microsoft Windows (10/11, PowerShell-capable endpoints); Google Chrome Browser (credential store targeted) |
| Malware | Satacom |
| MITRE ATT&CK techniques | T1027 Obfuscated Files or Information, T1055 Process Injection, T1059 Command and Scripting Interpreter, T1071 Application Layer Protocol, T1132 Data Encoding, T1140 Deobfuscate/Decode Files or Information, T1218 System Binary Proxy Execution, T1497 Virtualization/Sandbox Evasion, T1547 Boot or Logon Autostart Execution, T1555 Credentials from Password Stores, T1566 Phishing, T1573 Encrypted Channel and 3 more in the attack flow |
| Motivation | FINANCIAL |
| Target sectors | generalpublic, enterprise |
| Target regions | south korea |
| Detection rules | 9 (1 shown free) |
| Sources | 7 cited sources |
Detection coverage
As of 2026-09-11, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-2441 across Splunk SPL, Microsoft KQL and Sigma, covering 30 indicator(s) of compromise. The whole corpus is readable without an account. Every threat comes with one complete detection in Splunk SPL, Microsoft KQL and Sigma, free for everyone; a free account adds daily TL Creds to unlock the rest. Paid plans add a larger daily allowance, correlation and the MCP server. Threadlinqs MCP server · View plans.
Free detection: ClickFix Fake CAPTCHA PowerShell Clipboard Execution via Run Dialog (LegionLoader) — SPL
Detects the LegionLoader ClickFix exploitation chain: PowerShell spawned directly from explorer.exe (consistent with a Win+R Run-dialog paste) using a hidden window and execution-policy bypass, then fetching and running a next-stage script in-memory via Invoke-RestMethod/irm from a .monster or .cfd distribution domain. Matches victims lured through the fake Cloudflare CAPTCHA page (namu.wiki compromise or account-provisioning spear-phishing).
- Languages: Splunk SPL, Microsoft KQL, Sigma
- ATT&CK: T1566.002, T1204.004, T1059.001, T1583.001
- Data sources: sysmon, windows_security, microsoft_defender_for_endpoint
- Severity: high
- Confidence: high
Splunk SPL
index=sysmon sourcetype="XmlWinEventLog:Microsoft-Windows-Sysmon/Operational" EventCode=1
| eval cmdline=lower(CommandLine)
| where match(Image, "(?i)\\\\powershell\\.exe$") AND match(ParentImage, "(?i)\\\\explorer\\.exe$")
| where (like(cmdline, "%-w h%") OR like(cmdline, "%-windowstyle hidden%")) AND like(cmdline, "%-ep bypass%")
| where like(cmdline, "%invoke-restmethod%") OR like(cmdline, "%irm %")
| where match(cmdline, "\\.(monster|cfd)")
| table _time, ComputerName, User, ParentImage, Image, CommandLine
| sort -_timeMicrosoft KQL
DeviceProcessEvents
| where Timestamp > ago(24h)
| where FileName =~ "powershell.exe"
| where InitiatingProcessFileName =~ "explorer.exe"
| where ProcessCommandLine has_any ("-w h", "-windowstyle hidden")
| where ProcessCommandLine has "-ep bypass"
| where ProcessCommandLine has_any ("Invoke-RestMethod", "irm ")
| where ProcessCommandLine has_any (".monster", ".cfd")
| project Timestamp, DeviceName, AccountName, InitiatingProcessFileName, FileName, ProcessCommandLineSigma
title: ClickFix Fake CAPTCHA PowerShell Clipboard Execution via Run Dialog (LegionLoader)
id: 7f2a1c4e-9b3d-4e6a-8f21-3a5c7d9e1b02
status: experimental
description: Detects PowerShell launched from explorer.exe (Win+R Run dialog) with a hidden window, execution-policy bypass, and an in-memory download-and-run pattern to a .monster/.cfd domain, consistent with the ClickFix technique used to deliver LegionLoader via a fake Cloudflare CAPTCHA page.
references:
- https://asec.ahnlab.com/en/95374/
- https://attack.mitre.org/techniques/T1204/004/
author: Threadlinqs Team
date: 2026-09-11
tags:
- attack.initial-access
- attack.t1566.002
- attack.execution
- attack.t1059.001
- attack.t1204.004
- attack.tool.legionloader
logsource:
category: process_creation
product: windows
detection:
selection_parent:
ParentImage|endswith: '\explorer.exe'
Image|endswith: '\powershell.exe'
selection_flags:
CommandLine|contains:
- '-w h'
- '-windowstyle hidden'
selection_bypass:
CommandLine|contains: '-ep bypass'
selection_download:
CommandLine|contains:
- 'Invoke-RestMethod'
- 'irm '
selection_domain:
CommandLine|re: '\.(monster|cfd)'
condition: selection_parent and selection_flags and selection_bypass and selection_download and selection_domain
falsepositives:
- Legitimate IT/help-desk remote-support one-liners pasted into the Run dialog by administrators (rare; verify domain reputation and absence of AMSI bypass)
level: highKnown false positives:
- Legitimate IT/help-desk remote-support scripts pasted into the Run dialog by administrators (verify destination-domain reputation and absence of an AMSI-bypass flag before escalating)
Tuning notes: Extend the domain regex as new .monster/.cfd distribution domains rotate in — TL_OSINT_Scan flagged a low-confidence, community-observed sibling ClickFix campaign on triapfog[.]com worth adding to the watchlist pending confirmation. Tighten further by baselining how often ParentImage=explorer.exe PowerShell launches occur per host in the environment; Run-dialog-spawned PowerShell is rare outside admin workflows.
- ClickFix Fake CAPTCHA PowerShell Clipboard Execution via Run Dialog (LegionLoader) — KQL · unlock with TL Creds
- LegionLoader Process Hollowing, Registry Persistence, and AMSI-Bypass/Credential-Theft Activity — KQL · unlock with TL Creds
- LegionLoader RC4/Base64 HTTP C2 Beacon to nfront.php/nback.php Endpoints — KQL · unlock with TL Creds
- ClickFix Fake CAPTCHA PowerShell Clipboard Execution via Run Dialog (LegionLoader) — Sigma · unlock with TL Creds
- LegionLoader Process Hollowing, Registry Persistence, and AMSI-Bypass/Credential-Theft Activity — Sigma · unlock with TL Creds
- LegionLoader RC4/Base64 HTTP C2 Beacon to nfront.php/nback.php Endpoints — Sigma · unlock with TL Creds
- LegionLoader Process Hollowing, Registry Persistence, and AMSI-Bypass/Credential-Theft Activity — SPL · unlock with TL Creds
- LegionLoader RC4/Base64 HTTP C2 Beacon to nfront.php/nback.php Endpoints — SPL · unlock with TL Creds
Indicators of compromise
As of 2026-09-11, Threadlinqs Intelligence tracks 30 indicators of compromise for TL-2026-2441.
Sample indicators (3 of 30)
- IP address:
89.34.90[.]217 - Domain:
dallasbackstage[.]com - File hash:
0128a4b78aab83796042118ecb3e46e0
Defanged, TLP:CLEAR. Full set of 30 indicators: unlock with TL Creds.
Technical analysis
LegionLoader is a multi-stage loader/backdoor malware family, historically documented since December 2019 (SecurityWeek, Deep Instinct) as a malware-as-a-service dropper written in MS Visual C++ 8. In the campaign identified by AhnLab ASEC (published 2026-09-11), the malware is delivered through the ClickFix technique: a fake Cloudflare CAPTCHA verification screen instructs victims to copy a PowerShell one-liner and paste it into the Windows Run dialog or a command prompt, rather than performing a normal button-click or text-entry CAPTCHA challenge.
Two delivery vectors were confirmed. First, attackers inserted malicious links disguised as official content on a South Korean namu.wiki page (the 'Shinon Hope Town' article); visitors are carried through multiple redirect hops before landing on the fake CAPTCHA page. Second, attackers sent spear-phishing emails impersonating an internal account-provisioning system, using urgency (warnings that incorrect password entry would lock the account, and that the account must be activated within a deadline) to drive victims at specific target enterprises to the same fake CAPTCHA flow.
The clipboard payload executes 'powershell -w h -ep bypass -c' with an inline script downloaded via Invoke-RestMethod from a rotating '.monster'/'.cfd' distribution domain, parsed and run in-memory via the PowerShell Language Parser (no script ever touches disk). This retrieves and runs the LegionLoader stage, which sequentially decrypts embedded, encrypted shellcode and PE payloads and executes the final backdoor through a custom PE loader.
Before executing, the loader performs anti-analysis checks: it enumerates display devices via EnumDisplayDevicesA and compares them against 16 blacklisted virtualization/remote-desktop adapter strings (VMware, Hyper-V, VirtualBox, Parallels, Citrix, etc.), and queries ipinfo.io to determine whether its public IP belongs to a hosting-provider ASN (a common sandbox/VPS indicator). If these checks pass, the backdoor supports multiple execution primitives driven from its C2: native PE execution via process hollowing into a suspended RuntimeBroker.exe, shellcode injection into explorer.exe, PowerShell execution with an AMSI bypass, and MSI payload execution via ShellExecute/msiexec. For credential theft, it extracts the Chrome Legacy Master Key and App-Bound Encryption key along with browser profile/history data, enabling decryption of stored Chrome credentials and cookies offline. Persistence is established through Windows Registry Run and RunOnce keys.
C2 communication uses RC4-encrypted, Base64-encoded traffic over HTTP to two confirmed domains (dallasbackstage[.]com and gemscocl[.]com, resolving to 89.34.90.217), each exposing /nfront.php (host registration / command retrieval) and /nback.php (execution-result reporting) endpoints. Distribution infrastructure shows clear operational rotation: ten '.monster'-TLD domains were active in August 2026 (including leappoploaderinstantly[.]monster, the domain observed in the captured PowerShell command), with a brief, overlapping use of a '.com' domain (socialseoguru[.]com), followed by a shift to five '.cfd'-TLD domains in September 2026 — consistent with active adversary infrastructure cycling to evade domain-reputation blocking.
MITRE ATT&CK techniques by tactic:
Stealth
T1027 Obfuscated Files or Information; T1055 Process Injection; T1140 Deobfuscate/Decode Files or Information; T1497 Virtualization/Sandbox Evasion
Execution
T1059 Command and Scripting Interpreter
Command and Control
T1071 Application Layer Protocol; T1132 Data Encoding; T1573 Encrypted Channel
stealth
T1218 System Binary Proxy Execution
Persistence
T1547 Boot or Logon Autostart Execution
Credential Access
T1555 Credentials from Password Stores
Initial Access
Resource Development
Reconnaissance
T1589 Gather Victim Identity Information
defense-impairment
Remediation
Immediate actions
- Block the confirmed C2 domains (dallasbackstage[.]com, gemscocl[.]com) and IP (89.34.90.217) at DNS/perimeter firewall
- Block the confirmed distribution domains across .monster, .cfd, and .com TLDs listed in IOCs
- Hunt for the five listed MD5 hashes across EDR/AV telemetry
- Search PowerShell/AMSI script-block logs for 'irm' invocations against .monster or .cfd domains and for the '-w h -ep bypass' command pattern
- Alert on RuntimeBroker.exe processes created in a suspended state or exhibiting hollowed-memory characteristics
- Force Chrome credential/App-Bound key rotation and re-authentication for any user who ran the clipboard command
Workarounds
- Block execution of PowerShell from Explorer/Run-dialog paste actions via endpoint policy where feasible
- Filter/quarantine inbound email referencing account provisioning/activation with external links pending verification
Longer-term hardening
- Deploy EDR with behavioral detection for process hollowing (T1055.012), AMSI-bypass patterns, and PowerShell abuse
- Restrict PowerShell execution policy and enable Constrained Language Mode / AppLocker for non-admin users
- Deploy Group Policy or endpoint controls that disable the Windows Run dialog's ability to execute pasted multi-stage commands unattended, and monitor clipboard-to-terminal execution patterns (ClickFix detection)
- Run security-awareness training specifically covering ClickFix / fake-CAPTCHA 'verify you are human' prompts that ask users to run commands
Affected and fixed versions
Timeline
- Attackers briefly use a .com-TLD domain (locked-ioc-27) alongside the .monster infrastructure during the same period.
- LegionLoader ClickFix distribution campaign begins using a wave of ten newly registered .monster-TLD domains as first-stage staging infrastructure (August 2026).
- Distribution infrastructure rotates from .monster-TLD domains to a new set of five .cfd-TLD domains, consistent with active operational security by the operators (September 2026).
- AhnLab ASEC publishes an English-language advisory corroborating the same LegionLoader ClickFix findings for a broader audience.
- AhnLab ASEC publishes Korean-language advisory documenting the LegionLoader ClickFix campaign, including hashes, domains, and C2 IP.
- Spear-phishing emails impersonating an internal account-provisioning system, targeting specific enterprises, identified driving victims to the same fake CAPTCHA/ClickFix flow.
- Compromised namu.wiki 'Shinon Hope Town' page identified redirecting visitors through multiple hops to a fake Cloudflare CAPTCHA verification page.
Sources cited
- 보안 인증 창을 위장한 LegionLoader 악성코드 (ClickFix 기법으로 유포) asec.ahnlab.com · Vendor report · 2026-09-11
- I just trusted the security certificate prompt... Beware of the LegionLoader malware being distributed via the ClickFix method asec.ahnlab.com · Vendor report · 2026-09-11
- User Execution: Malicious Copy and Paste, Sub-technique T1204.004 MITRE · Research · 2025-03-18
- Process Injection: Process Hollowing, Sub-technique T1055.012 MITRE · Research · 2026-05-12
- Untangling Legion Loader's Hornet Nest of Malware deepinstinct.com · Blog · 2019-12-18
- New Legion Loader Delivers a Variety of Malware SecurityWeek · News · 2019-12-19
- ClickFix Attacks Expand Using Fake CAPTCHAs, Microsoft Scripts, and Trusted Web Services The Hacker News · News · 2026-01-27
Previous and next in malware
Drafted by the Threadlinqs research pipeline from 7 sources; published ; last updated —.