Threat reportPhishingTL-2026-2919
Milk Dragon (NaiLong) Phishing-as-a-Service Kit Uses Facebook and TikTok Discounts to Steal Cards and Bypass MFA
Milk Dragon (aka NaiLong) is a Telegram-sold phishing-as-a-service kit that lures victims with discount-themed Facebook and TikTok posts to counterfeit WordPress/WooCommerce storefronts, harvesting payment card data and relaying 3-D Secure/MFA codes in real time via adversary-in-the-middle bank templates. Group-IB linked 258 phishing pages since October 2025, with victims in 66 countries.
Key facts
| Identifiers | TL-2026-2919 |
|---|---|
| Also known as | Milk Dragon Phishing Kit, NaiLong Phishing Kit |
| Severity | HIGH |
| Affected products | WordPress WooCommerce checkout (abused by attacker-controlled storefronts, not a vulnerability); Consumers and retail brands Online shoppers and impersonated brands (e.g. LEGO, Calvin Klein, Aeon Malaysia) |
| Attribution | Milk Dragon (low confidence) |
| Malware | Milk Dragon, NaiLong, telegram |
| Tools | WooCommerce |
| MITRE ATT&CK techniques | T1056.001 Keylogging, T1056.003 Web Portal Capture, T1071.001 Web Protocols, T1111 Multi-Factor Authentication Interception, T1204.001 Malicious Link, T1557 Adversary-in-the-Middle, T1566.003 Spearphishing via Service, T1583.001 Domains, T1585.001 Social Media Accounts, T1657 Financial Theft, T1684.001 Impersonation |
| Motivation | FINANCIAL |
| Target sectors | retail, ecommerce, consumer goods, financial services, supermarkets |
| Target regions | Global |
| Detection rules | 9 (1 shown free) |
| Sources | 6 cited sources |
Detection coverage
As of 2026-10-04, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-2919 across Splunk SPL, Microsoft KQL and Sigma, covering 14 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: Milk Dragon — Facebook/TikTok Discount Lure Click-Through to WooCommerce Checkout on Unapproved Storefront
Detects web-proxy requests where the HTTP referrer is Facebook or TikTok and the destination is a WooCommerce-style cart/checkout path (/checkout, /cart, ?wc-ajax=) on a domain outside the approved-shop allowlist. This is the Milk Dragon (NaiLong) distribution pattern: native-looking discount posts on social media that send victims to counterfeit WordPress/WooCommerce storefronts.
- Languages: Splunk SPL, Microsoft KQL, Sigma
- ATT&CK: T1566.003, T1204.001, T1656
- Data sources: web_proxy:zscalernss-web, web_proxy:squid:access, CommonSecurityLog, sigma:proxy
- Severity: medium
- Confidence: medium
Splunk SPL
```Milk Dragon: social referral landing on a WooCommerce checkout of an unapproved shop```
index=proxy sourcetype IN ("zscalernss-web","bluecoat:proxysg:access:syslog","squid:access") earliest=-24h
| eval ref=lower(coalesce(http_referrer, referrer, "")), shop_host=lower(coalesce(dest_host, url_domain, dest, "")), upath=lower(coalesce(uri_path, "")), uq=lower(coalesce(uri_query, ""))
| where match(ref, "(facebook[.]com|fb[.]com|tiktok[.]com)")
| where match(upath, "/(checkout|cart)") OR like(uq, "%wc-ajax=%")
```exclude social platforms and well-known marketplaces; extend with your approved_shop_domains list```
| where NOT match(shop_host, "(^|[.])(facebook[.]com|fb[.]com|instagram[.]com|tiktok[.]com|amazon[.]com|ebay[.]com|etsy[.]com|shopify[.]com)$")
| stats count AS requests dc(upath) AS distinct_paths values(upath) AS paths min(_time) AS first_seen max(_time) AS last_seen BY src, user, shop_host
| convert ctime(first_seen) ctime(last_seen)
| sort - requestsMicrosoft KQL
// Milk Dragon: social referral landing on a WooCommerce checkout of an unapproved shop
CommonSecurityLog
| where TimeGenerated > ago(24h)
| where isnotempty(RequestURL)
| where RequestContext has_any ("facebook.com", "fb.com", "tiktok.com")
| where RequestURL has "/checkout" or RequestURL has "/cart" or RequestURL has "wc-ajax="
| extend ShopHost = tolower(coalesce(DestinationHostName, tostring(parse_url(RequestURL).Host)))
| where not(ShopHost matches regex @"(^|[.])(facebook[.]com|fb[.]com|instagram[.]com|tiktok[.]com|amazon[.]com|ebay[.]com|etsy[.]com|shopify[.]com)$")
| summarize Requests = count(), Urls = make_set(RequestURL, 10), FirstSeen = min(TimeGenerated), LastSeen = max(TimeGenerated) by SourceIP, SourceUserName, ShopHost
| order by Requests descSigma
title: Milk Dragon Social Media Discount Lure Click-Through to WooCommerce Checkout
id: 7c1e5a42-93b8-4f0d-a6d1-2b8e4c9f1a01
status: experimental
description: Detects proxy requests to a WooCommerce cart/checkout path where the referrer is Facebook or TikTok, the distribution pattern of the Milk Dragon (NaiLong) phishing-as-a-service kit.
references:
- https://www.group-ib.com/blog/milk-dragon-nailong-phishing-kit/
author: Threadlinqs Team
date: 2026/10/04
tags:
- attack.initial-access
- attack.t1566.003
- attack.execution
- attack.t1204.001
- attack.defense-evasion
- attack.t1656
logsource:
category: proxy
detection:
selection_referrer:
cs-referrer|contains:
- 'facebook.com'
- 'fb.com'
- 'tiktok.com'
selection_checkout:
c-uri|contains:
- '/checkout'
- '/cart'
- 'wc-ajax='
filter_platforms:
cs-host|endswith:
- '.facebook.com'
- '.fb.com'
- '.tiktok.com'
- '.instagram.com'
- '.amazon.com'
- '.ebay.com'
condition: selection_referrer and selection_checkout and not filter_platforms
falsepositives:
- Legitimate merchants that run Facebook or TikTok ad campaigns landing on their own checkout
- Employees shopping on approved retail sites
level: mediumKnown false positives:
- Legitimate merchants running Facebook/TikTok ad campaigns that land on their own checkout
- Employees purchasing from approved retailers via social ads
Tuning notes: Maintain an approved_shop_domains lookup and exclude it; keep the referrer filter. Requires proxy logs with the referrer field (CEF requestContext for KQL). Prioritise hits where the shop domain was first seen in the last 30 days.
- Milk Dragon — Brand-Impersonating Discount Storefront Domains (LEGO, Calvin Klein, Aeon) · unlock with TL Creds
- Milk Dragon — [indicator] / [indicator] Files Written Under wp-content/plugins on Web Hosts · unlock with TL Creds
- Milk Dragon — Web Server / PHP Process Connecting to New External Operator Backend (API Base URL) · unlock with TL Creds
- Milk Dragon — Spoofed 3-D Secure OTP / Turnstile Interstitial on a WooCommerce Storefront (AiTM OTP Relay) · unlock with TL Creds
- Milk Dragon — Browser Fetches [indicator] Plugin Assets From a Counterfeit Storefront · unlock with TL Creds
- Milk Dragon — Web Server / PHP Process Contacting Telegram Bot API (Operator Alert Channel) · unlock with TL Creds
- Milk Dragon — [indicator] / [indicator] Install or Activation on WordPress via wp-admin Access Logs · unlock with TL Creds
- Milk Dragon — Persistent [indicator] WebSocket Session From Storefront (Per-Keystroke Exfiltration Channel) · unlock with TL Creds
Indicators of compromise
As of 2026-10-04, Threadlinqs Intelligence tracks 14 indicators of compromise for TL-2026-2919.
Technical analysis
Group-IB (report dated 2026-10-01, covered by GBHackers on 2026-10-02) documents Milk Dragon, also tracked as NaiLong, a phishing-as-a-service (PhaaS) operation active since at least October 2025. The kit is marketed in Telegram communities from 300 USDT per month and includes access to an operator panel, updates, support and role-based accounts. Group-IB identified 258 phishing pages tied to the kit, with victims across 66 countries and 21 impersonated brands spanning cosmetics, fashion, food and beverages, home and baby products, toys and regional supermarket chains (examples named: LEGO, Calvin Klein, Aeon Malaysia).
Distribution relies on native-looking social media posts on Facebook and TikTok that advertise heavily discounted products, supported by fake profiles with AI-generated content and inflated follower counts. A FOMO/discount lure replaces the fear-based pretext of traditional phishing. Clicking through lands the victim on a counterfeit WordPress storefront built on the legitimate WooCommerce plugin.
The core component is a custom malicious WordPress plugin named BytePress (also referenced as the 'SP plugin'). It adds fraudulent payment options (credit card and PayPal) to the WooCommerce checkout, exposes an 'API Base URL' setting that links the site to the operator backend, and holds a persistent Socket.IO WebSocket connection to the operator's server. Victim input is streamed character by character before the form is ever submitted. After card entry the victim sees a fake turnstile/loading screen, then a spoofed 3-D Secure one-time-password page selected from 36 financial-institution templates; the operator can redirect the victim to a different OTP method in real time and relay the code to the legitimate payment processor (adversary-in-the-middle), followed by a fake order-confirmation page.
The operator panel supports a many-to-one multi-domain architecture, role-based subordinate accounts, visitor/order/conversion metrics, card BIN identification with automatic bank tagging, live session monitoring and control, Telegram bot alerts for new submissions, a custom template builder, a central SQL database of victim data with re-targeting profiles, and Docker-containerized deployment with automated setup. Group-IB restricts the full IOC list (domains, IPs, Telegram handles) to its Threat Intelligence portal customers, so no network indicators are publicly disclosed.
MITRE ATT&CK techniques by tactic:
Collection
Credential Access
T1056.003 Web Portal Capture; T1111 Multi-Factor Authentication Interception; T1557 Adversary-in-the-Middle
Command and Control
Execution
Initial Access
T1566.003 Spearphishing via Service
Resource Development
T1583.001 Domains; T1585.001 Social Media Accounts
Impact
Defense Evasion
Remediation
Immediate actions
- Monitor and take down lookalike retail/storefront domains impersonating your brand
- Report fraudulent Facebook and TikTok posts and fake merchant profiles to the platforms
- Flag card-not-present transactions with anomalous checkout patterns and rapid 3-D Secure OTP retries
- Advise customers to notify their bank immediately if card data or OTPs were entered on a suspicious shop
Workarounds
- Hunt for unexpected WordPress plugins (e.g. BytePress) or WooCommerce payment gateways with an 'API Base URL' setting on sites you operate
- Inspect for persistent Socket.IO/WebSocket connections from checkout pages to unfamiliar hosts
Longer-term hardening
- Move payment authentication toward phishing-resistant methods (app-based transaction signing, passkeys) rather than relayable SMS/OTP codes
- Integrate threat intelligence on PhaaS kits and brand-impersonation monitoring
- Educate users that steep social-media discounts are a deception indicator and to verify merchants via urlscan.io, VirusTotal or ScamAdviser
Affected and fixed versions
- WordPress — WooCommerce checkout (abused by attacker-controlled storefronts, not a vulnerability)
- Consumers and retail brands — Online shoppers and impersonated brands (e.g. LEGO, Calvin Klein, Aeon Malaysia)
Timeline
- Earliest activity linked to the Milk Dragon (NaiLong) kit; Group-IB dates the operation to October 2025 (exact day not published, first of month used).
- Group-IB documents the Telegram-sold PhaaS model from 300 USDT/month with operator panel, role-based accounts, BIN tagging, Telegram alerts and Docker deployment.
- Group-IB documents 36 financial-institution 3-D Secure/OTP templates and real-time operator-controlled OTP relay (AiTM).
- Group-IB details the BytePress WordPress plugin: fake WooCommerce card/PayPal payment options, 'API Base URL' operator link, Socket.IO character-by-character keystroke streaming.
- Group-IB publishes its analysis: 258 phishing pages tied to the kit since October 2025, victims in 66 countries, 21 impersonated brands.
- GBHackers on Security reports the Group-IB findings, bringing the campaign to wider defender attention.
- Threadlinqs opens TL-2026-2919 to track the campaign; no public network IOCs available (Group-IB restricts them to portal customers).
Sources cited
- Group-IB: Milk Dragon (NaiLong) Phishing Kit group-ib.com · Research · 2026-10-01
- MITRE ATT&CK T1557 Adversary-in-the-Middle MITRE · Research · 2026-10-04
- MITRE ATT&CK T1111 Multi-Factor Authentication Interception MITRE · Research · 2026-10-04
- MITRE ATT&CK T1566.003 Spearphishing via Service MITRE · Research · 2026-10-04
- MITRE ATT&CK T1056.003 Web Portal Capture MITRE · Research · 2026-10-04
- GBHackers: Milk Dragon Phishing Kit Uses Facebook and TikTok Discounts to Steal Cards and Bypass MFA gbhackers.com · News · 2026-10-02
Previous and next in phishing
Drafted by the Threadlinqs research pipeline from 6 sources; published ; last updated —.