Skip to navigation

Threat reportMalwareTL-2026-6565

October 2026 Infostealer Trend Report: Lzveil, StealC, Remus, DarkCloud, AgentTesla

mediumACTIVE

AhnLab ASEC's October 2026 infostealer trend report describes a newly named infostealer, Lzveil, delivered together with the Remus stealer inside a 7Z archive that contains a password-protected RAR. StealC is also seen in crack/keygen lures spread through SEO poisoning and file hosting, while DarkCloud and AgentTesla arrive as phishing email attachments.

Key facts

IdentifiersTL-2026-6565
SeverityMEDIUM
First disclosed
Affected productsMicrosoft Windows (user endpoints running pirated software or opening phishing attachments)
MalwareAgent Tesla, DarkCloud, Lzveil, Mega, Remus, Stealc
MITRE ATT&CK techniquesT1005 Data from Local System, T1027 Obfuscated Files or Information, T1036 Masquerading, T1048.003 Exfiltration Over Unencrypted Non-C2 Protocol, T1055 Process Injection, T1056.001 Keylogging, T1071.001 Web Protocols, T1113 Screen Capture, T1204.002 Malicious File, T1539 Steal Web Session Cookie, T1555.003 Credentials from Web Browsers, T1566.001 Spearphishing Attachment and 2 more in the attack flow
MotivationFINANCIAL
Target sectorsconsumer, technology, logistics, manufacturing
Target regionsGlobal, south korea
Detection rules9 (1 shown free)
Sources9 cited sources

Detection coverage

As of 2026-11-20, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-6565 across Splunk SPL, Microsoft KQL and Sigma, covering 20 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: Infostealer python37.dll DLL Side-Loading From User-Writable Path — Exploitation

Detects python37.dll (the main malicious DLL in the ~1.6% of ASEC-analysed crack/keygen infostealer samples that use DLL side-loading) loaded by a non-Python host executable running from a user-writable directory. Covers Lzveil, StealC and Remus lures (T1574.002 DLL Side-Loading, T1036 Masquerading) in TL-2026-6565.

  • Languages: Splunk SPL, Microsoft KQL, Sigma
  • ATT&CK: T1574.002, T1036
  • Data sources: sysmon:EventCode=7 (index=sysmon, XmlWinEventLog:Microsoft-Windows-Sysmon/Operational), MDE:DeviceImageLoadEvents, sigma:windows/image_load
  • Severity: high
  • Confidence: high

Splunk SPL

index=sysmon sourcetype="XmlWinEventLog:Microsoft-Windows-Sysmon/Operational" EventCode=7 ImageLoaded="*\\python37.dll" earliest=-24h
    NOT Image IN ("*\\python.exe", "*\\pythonw.exe", "*\\py.exe", "*\\pyw.exe")
    Image IN ("*\\Users\\*", "*\\ProgramData\\*", "*\\Temp\\*", "*\\Downloads\\*")
```Behaviour: python37.dll loaded by a non-Python host running from a user-writable path. ASEC reports python37.dll as the main side-loaded DLL for crack/keygen infostealers such as Lzveil, StealC and Remus```
| eval dll_unsigned=if(lower(Signed)="false", "yes", "no")
| stats count min(_time) as first_seen max(_time) as last_seen values(ImageLoaded) as loaded_dll values(dll_unsigned) as dll_unsigned values(SignatureStatus) as dll_signature_status values(Hashes) as hashes by host, User, Image
| eval first_seen=strftime(first_seen, "%Y-%m-%d %H:%M:%S"), last_seen=strftime(last_seen, "%Y-%m-%d %H:%M:%S")
| table host User Image loaded_dll dll_unsigned dll_signature_status hashes count first_seen last_seen

Microsoft KQL

// TL-2026-6565 python37.dll side-loading by non-Python host in user-writable path (T1574.002)
DeviceImageLoadEvents
| where Timestamp > ago(24h)
| where FileName =~ "python37.dll"
| where not(InitiatingProcessFileName in~ ("python.exe", "pythonw.exe", "py.exe", "pyw.exe"))
| where InitiatingProcessFolderPath has_any (@"\Users\", @"\ProgramData\", @"\Temp\", @"\Downloads\")
| extend SameDirSideload = tolower(replace_string(FolderPath, FileName, "")) == tolower(replace_string(InitiatingProcessFolderPath, InitiatingProcessFileName, ""))
| project Timestamp, DeviceName, Account = InitiatingProcessAccountName, HostProcess = InitiatingProcessFileName, HostPath = InitiatingProcessFolderPath, DllPath = FolderPath, SameDirSideload, DllSHA1 = SHA1, HostSHA256 = InitiatingProcessSHA256
| order by Timestamp desc

Sigma

title: Python37.dll Side-Loaded by Non-Python Host From User-Writable Path
id: 7c1f0a52-3b8e-4d6a-9f21-5a4e8c0b1d11
status: experimental
description: 'Detects python37.dll loaded by a non-Python executable running from a user-writable directory. ASEC October 2026 infostealer report names python37.dll as the main malicious DLL in side-loading infostealer samples (Lzveil, StealC, Remus lures). TL-2026-6565.'
references:
    - https://asec.ahnlab.com/en/96698/
    - https://www.gendigital.com/blog/insights/research/remus-64bit-variant-of-lumma-stealer
author: Threadlinqs Team
date: 2026-11-20
tags:
    - attack.defense-evasion
    - attack.persistence
    - attack.t1574.002
    - attack.t1036
logsource:
    category: image_load
    product: windows
detection:
    selection:
        ImageLoaded|endswith: '\python37.dll'
    selection_path:
        Image|contains:
            - '\Users\'
            - '\ProgramData\'
            - '\Temp\'
            - '\Downloads\'
    filter_python:
        Image|endswith:
            - '\python.exe'
            - '\pythonw.exe'
            - '\py.exe'
            - '\pyw.exe'
    condition: selection and selection_path and not filter_python
falsepositives:
    - Portable applications or games that legitimately bundle an embedded Python 3.7 runtime
    - Developer tools run from user profile directories
level: high

Known false positives:

  • Portable apps or games bundling an embedded Python 3.7 runtime and a renamed launcher
  • Developer tooling unpacked into user profile directories

Tuning notes: Allowlist known-good host executables by hash or signer after a 14-day baseline. Raise priority when dll_unsigned=yes or SameDirSideload is true. Ensure Sysmon EventCode 7 is enabled for DLLs loaded from user paths.

8 more detection rules (Splunk SPL, Microsoft KQL, Sigma) · unlock with TL Creds. Sign up free to unlock the detection rules or query them with the Threadlinqs MCP server (Purple and Gold in full, 2 calls a day on a free account)
  • Crack/Keygen Infostealer Lure Executed From Archive Temp Directory — Exploitation · unlock with TL Creds
  • .NET Utility Spawned as Injection Host by User-Path Loader — Post-Exploitation · unlock with TL Creds
  • Remus EtherHiding RPC and Infostealer IOC Domain Resolution — C2 · unlock with TL Creds
  • Infostealer Phishing Attachment Executed From Mail Client Path — Exploitation · unlock with TL Creds
  • ASEC Infostealer Hash and ClickFix Domain Execution Match — Post-Exploitation · unlock with TL Creds
  • Remus C2 Over Non-Standard TCP 5902 From User-Path Binary — C2 · unlock with TL Creds
  • Browser Credential and Cookie Store Staging by Non-Browser Process — Post-Exploitation · unlock with TL Creds
  • Infostealer SMTP and FTP Exfiltration From User-Path Binary — C2 / Exfiltration · unlock with TL Creds

Indicators of compromise

As of 2026-11-20, Threadlinqs Intelligence tracks 20 indicators of compromise for TL-2026-6565.

Sample indicators (3 of 20)

  • Domain: nzqztr2208262ea[.]cfd
  • File hash: 0094855366c30a9da7d6f4962cb2206f
  • Domain: arkedfile[.]s3[.]us-east-1[.]amazonaws[.]com

Defanged, TLP:CLEAR. Full set of 20 indicators: unlock with TL Creds.

20 indicators of compromise · unlock with TL Creds. Compare plans

Community OSINT corroboration

1 of this threat's indicators have also been reported by the open-source security community. Community sightings are unverified and are kept separate from Threadlinqs' curated indicators. Indicator values, reporters and campaign linkage are available to authenticated Red-tier users.

Technical analysis

AhnLab ASEC's monthly infostealer trend report for October 2026 (published 2026-11-20) is built from ASEC's automated collection system, email honeypot, C2 automated analysis system and AhnLab product diagnostic logs. It reports that infostealers are mainly distributed disguised as cracks and keygens. Lures are promoted through SEO poisoning (search-result ranking manipulation) and hosted on file-sharing and cloud services including Mega and MediaFire, plus an Amazon S3 bucket and a '.cfd' domain. Victims who search for pirated software download an archive and run the payload themselves.

The report's headline finding is Lzveil, described as a new infostealer. Lzveil was distributed together with Remus: the files sit in a password-protected RAR nested inside a 7Z archive, and both stealers were observed executing at the same time on the victim host. StealC is also named as a family frequently identified in the crack/keygen channel. Of the samples ASEC analysed, about 98.4% run as plain EXE files and about 1.6% use DLL side-loading, with python37.dll as the main malicious DLL. ASEC notes that many samples carry version and certificate metadata impersonating Microsoft Corporation (most frequent), Stashabuz Ltd., Tuplemix Technologies and Snako, Inc.; those companies are not involved.

Phishing is the second channel. A DarkCloud sample arrived as an attachment in an email impersonating a US laboratory-equipment company. It collects documents, keylogging data, email client data, browser data, screenshots and cryptocurrency wallet data, and exfiltrated over SMTP in the analysed sample. An AgentTesla sample, a .NET information stealer, arrived in an email impersonating a shipping company and exfiltrated over FTP.

Background on Remus from other public research (not stated in the ASEC report): Gen Threat Labs (2026-04-07) and Flashpoint (2026-07-02) describe Remus as a 64-bit stealer closely related to Lumma, reached via a test-build family called Tenzor, that resolves C2 using EtherHiding and decrypts an embedded ChaCha20 C2 config. Flashpoint reports it sold as MaaS in three tiers (Basic $250, Pro $500, Enterprise $1,000). CIS (MS-ISAC) reports delivery via ClickFix/PLYCHIP, DonutLoader and cracked-software installers, HTTP C2 on non-standard ports, session-cookie theft aimed at MFA bypass, and, from a 29 July developer post, targeting of AI coding clients. The ASEC report does not state a CVE, a threat actor or the C2 servers of Lzveil, so none are asserted here. ASEC does not say which family each listed MD5 belongs to. Severity is an analyst estimate.

MITRE ATT&CK techniques by tactic:

Collection

T1005 Data from Local System; T1056.001 Keylogging; T1113 Screen Capture

Defense Evasion

T1027 Obfuscated Files or Information; T1036 Masquerading; T1055 Process Injection; T1574.001 DLL

Exfiltration

T1048.003 Exfiltration Over Unencrypted Non-C2 Protocol

Command and Control

T1071.001 Web Protocols

Execution

T1204.002 Malicious File

Credential Access

T1539 Steal Web Session Cookie; T1555.003 Credentials from Web Browsers

Initial Access

T1566.001 Spearphishing Attachment

Resource Development

T1608.006 SEO Poisoning

Remediation

Immediate actions

  • Hunt for 7Z archives containing password-protected RAR files executed from user Downloads/Temp paths, and for python37.dll loaded by non-Python host executables
  • Block or alert on the distribution domain nzqztr2208262ea.cfd and review proxy logs for the arkedfile S3 bucket URL
  • Search endpoints for the five MD5 hashes listed in the IOC set
  • Reset credentials, invalidate browser sessions and rotate crypto-wallet keys for any host where an infostealer is confirmed

Workarounds

  • Do not block mediafire.com or mega.nz wholesale; instead inspect downloads from them and block archive/executable downloads for non-approved users

Longer-term hardening

  • Prohibit installation of cracked software and keygens; restrict execution from user-writable download locations
  • Enforce MFA and discourage saving passwords in browsers; deploy EDR with detection for browser credential-store access and DLL side-loading
  • Filter inbound email attachments from impersonated lab-equipment and shipping senders; alert on outbound SMTP/FTP from non-mail hosts

Affected and fixed versions

  • Microsoft — Windows (user endpoints running pirated software or opening phishing attachments)

Timeline

  • Gen Threat Labs: Tenzor test builds, the Lumma-to-Remus transitional family, carry an encrypted-string build date of 16 September 2025.
  • Gen Threat Labs places the first Remus campaigns at the turn of January/February 2026 (month-level precision).
  • Flashpoint first sees Remus offered for sale as a MaaS in March 2026 (month-level precision); CIS tracks distribution from March through September 2026.
  • Gen Threat Labs publishes its analysis of Remus as a 64-bit variant of Lumma Stealer using EtherHiding and a ChaCha20-encrypted C2 config.
  • Flashpoint publishes 'Remus Stealer: A New, Not-So-New Infostealer', detailing Basic/Pro/Enterprise tiers ($250/$500/$1,000).
  • Per CIS, the Remus developer announces the stealer now targets AI coding clients (Claude, Codex, OpenCode, Cursor, Devin).
  • End of the October 2026 observation period in which ASEC observed Lzveil co-delivered with Remus, StealC in crack/keygen lures, and DarkCloud and AgentTesla phishing campaigns.
  • AhnLab ASEC publishes the October 2026 Infostealer Trend Report naming Lzveil as a new infostealer and listing distribution domains and MD5 hashes.

Sources cited

Drafted by the Threadlinqs research pipeline from 9 sources; published ; last updated —.

Threadlinqs Intelligence — Real-Time Threat Detection Platform

// threat_feed
$ sort --newest
Showing all threats

Live intelligence console

Threat level

Latest Threats