Threat reportSupply ChainTL-2026-6186
Poisoned third-party scripts (Optimizely, Kissmetrics, StorifyMe, minute-ly) on BBC, Zoom and other major sites deliver financial-theft skimmer and iPhone exploit-loader router (Coruna / DarkSword-linked)
Hudson Rock reports malicious code injected into third-party scripts (Optimizely projectJS, Kissmetrics, StorifyMe, minute-ly) served on at least 21 named sites including BBC, Zoom, GitLab, HSBC UAE, CNN and Fox News. The code loads a browser financial-theft payload (track.js: crypto-address substitution, card skimming, conditional cookie/screenshot collection) and an iOS user-agent router that serves entry pages associated with Coruna and DarkSword exploit chains.
Key facts
| Identifiers | TL-2026-6186 |
|---|---|
| Also known as | Poisoned Optimizely projectJS, track.js skimmer, JIUJIU888, Caishen iOS C2 V2 |
| Severity | HIGH |
| First disclosed | |
| Affected products | Optimizely Customer projectJS (cdn.optimizely.com); Kissmetrics i.kissmetrics.com/i.js, i.kissmetrics.io/i.js, scripts.kissmetrics.io; StorifyMe storifyme-elements.min.js and custom widgets; minute-ly apv-static.minute-ly.com loader.js; Apple iOS (targeted by routed exploit-kit entries) |
| Weaknesses (CWE) | CWE-829, CWE-494, CWE-79 |
| Malware | Coruna, DarkSword |
| MITRE ATT&CK techniques | T1027 Obfuscated Files or Information, T1041 Exfiltration Over C2 Channel, T1056 Input Capture, T1059.007 JavaScript, T1071.001 Web Protocols, T1113 Screen Capture, T1115 Clipboard Data, T1140 Deobfuscate/Decode Files or Information, T1185 Browser Session Hijacking, T1189 Drive-by Compromise, T1190 Exploit Public-Facing Application, T1195.002 Compromise Software Supply Chain and 9 more in the attack flow |
| Motivation | FINANCIAL |
| Target sectors | news - media, technology, finance, telecoms, retail, insurance, education, government administration |
| Target regions | Global, united kingdom, united states of america, Europe, Middle East, australia |
| Detection rules | 9 (1 shown free) |
| Sources | 7 cited sources |
Detection coverage
As of 2026-11-17, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-6186 across Splunk SPL, Microsoft KQL and Sigma, covering 49 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: TL-2026-6186 Payment Card Exfiltration to 1688360.cc card.php - Post-Exploitation Collection
Detects requests to 1688360.cc/card.php, the endpoint receiving skimmed payment-card fields (n, e, c, u) from the track.js card hook, which captures form submissions and fetch/XHR POST bodies and uses sendBeacon when available. A hit means a browser on your network ran the skimmer on a page where card data may have been entered (T1056, T1657, T1041). Transmission of real card data was not demonstrated in the source reporting, so confirm with proxy body logs or payment-page review.
- Languages: Splunk SPL, Microsoft KQL, Sigma
- ATT&CK: T1056, T1657, T1041
- Data sources: proxy, web_gateway, defender_for_endpoint
- Severity: critical
- Confidence: high
Splunk SPL
index=proxy sourcetype IN ("zscaler:nss:web","bluecoat:proxysg:access:syslog","squid:access","stream:http") earliest=-24h
| eval dom=lower(coalesce(url_domain, dest_host, dest)), u=lower(coalesce(url, uri_path, "")), method=upper(coalesce(http_method, "UNKNOWN"))
| where dom="1688360.cc" AND match(u,"/card[.]php")
| stats count min(_time) as first_seen max(_time) as last_seen sum(bytes_out) as bytes_out values(method) as methods values(http_referrer) as origin_pages by src, user, dom
| eval first_seen=strftime(first_seen,"%F %T"), last_seen=strftime(last_seen,"%F %T")
| table first_seen last_seen src user dom methods origin_pages bytes_out countMicrosoft KQL
let cardHost = '1688360.cc';
let endpointHits = DeviceNetworkEvents
| where Timestamp > ago(24h)
| where RemoteUrl has cardHost and RemoteUrl has '/card.php'
| project Timestamp, Source = 'endpoint', Device = DeviceName, Account = InitiatingProcessAccountName, Process = InitiatingProcessFileName, Method = '', Url = RemoteUrl;
let proxyHits = CommonSecurityLog
| where TimeGenerated > ago(24h)
| where RequestURL has cardHost and RequestURL has '/card.php'
| project Timestamp = TimeGenerated, Source = 'proxy', Device = SourceIP, Account = SourceUserName, Process = RequestClientApplication, Method = RequestMethod, Url = RequestURL;
union endpointHits, proxyHits
| summarize FirstSeen = min(Timestamp), LastSeen = max(Timestamp), Hits = count(), Methods = make_set(Method), Urls = make_set(Url, 10) by Source, Device, Account, Process
| order by LastSeen descSigma
title: TL-2026-6186 Payment Card Exfiltration to 1688360.cc card.php
id: 2d9e6b84-5a17-4c3f-8e02-91f4b7a3d602
status: experimental
description: 'Detects requests to the 1688360.cc card.php endpoint used by the track.js skimmer to send card number, expiry, CVV and page URL (fields n, e, c, u) from compromised web pages (TL-2026-6186).'
references:
- https://www.infostealers.com/article/poisoned-scripts-on-bbc-zoom-and-other-major-websites-linked-to-iphone-exploit-loaders/
- https://cloud.google.com/blog/topics/threat-intelligence/coruna-powerful-ios-exploit-kit
author: Threadlinqs Team
date: 2026-11-16
tags:
- attack.collection
- attack.t1056
- attack.exfiltration
- attack.t1041
- attack.impact
- attack.t1657
logsource:
category: proxy
detection:
selection:
cs-host: '1688360.cc'
c-uri|contains: '/card.php'
condition: selection
falsepositives:
- Researchers replaying the endpoint
- Sandbox detonation of the known sample
level: critical
Known false positives:
- Security researchers or sandboxes replaying the known skimmer endpoint
- Internal red-team or web-skimmer simulation exercises
Tuning notes: Any hit is high priority: identify the user, the origin page (http_referrer) and whether a payment form was involved. If your proxy logs POST bodies, check for n/e/c/u fields. Escalate to fraud and PCI teams if the page processes card data. Also review web-server and CSP report logs for connect-src violations to this host.
- TL-2026-6186 iOS Exploit-Kit Router Entry Pages and [indicator] on .icu Hosts - Exploitation · unlock with TL Creds
- TL-2026-6186 Cookie, URL and Screenshot Collection to [indicator] hit.php - Post-Exploitation Collection · unlock with TL Creds
- TL-2026-6186 iOS Router Gate and Telemetry API Calls (gs_state, pe_ok, chain_log) - C2 · unlock with TL Creds
- TL-2026-6186 Fetch of Known-Poisoned Optimizely and minute-ly Script Assets or Matching SHA256 - Exploitation · unlock with TL Creds
- TL-2026-6186 Attacker Wallet Addresses From track.js Clipper Seen in Web Traffic - Post-Exploitation · unlock with TL Creds
- TL-2026-6186 iOS User-Agent Fetch of JIUJIU888 Exploit-Chain Stages and frame.html - C2 and Delivery · unlock with TL Creds
- TL-2026-6186 Poisoned Third-Party Script Loads [indicator] track.js Skimmer - Exploitation · unlock with TL Creds
- TL-2026-6186 DNS Resolution of Skimmer and iOS Router C2 Infrastructure - C2 · unlock with TL Creds
Indicators of compromise
As of 2026-11-17, Threadlinqs Intelligence tracks 49 indicators of compromise for TL-2026-6186.
Sample indicators (3 of 49)
- IP address:
136.113.106[.]136 - Domain:
1688360[.]cc - URL:
hxxps://cdn[.]optimizely[.]com/public/4621041136/s/bbcx_prod.js
Defanged, TLP:CLEAR. Full set of 49 indicators: unlock with TL Creds.
Technical analysis
Successful device exploitation, zero-day use, operator identity, initial compromise mechanism and victim counts are not established.
Between 2026-11-13 and 2026-11-16 (UTC), archived and live responses from four third-party script providers were observed carrying injected code: Optimizely customer projectJS files (cdn.optimizely.com/js/<id>.js and /public/<id>/s/<name>.js), Kissmetrics (i.kissmetrics.com/i.js, i.kissmetrics.io/i.js, scripts.kissmetrics.io), StorifyMe (storifyme-elements.min.js and custom widgets on storifyme.com/.xyz) and minute-ly (apv-static.minute-ly.com/resources/v2/loader.js?v=3). Hudson Rock names 21 affected properties as a documented subset (not a victim count): bbc.com, zoom.com, about.gitlab.com, hsbc.ae, hootsuite.com, talkwalker.com, virginmediao2.co.uk, stamps.com, tricentis.com, securitybank.com, buzzstream.com, chicagoitsupport.com, faz.net, smh.com.au, unibas.ch, vidaxl.co.uk, salesforce.com, edition.cnn.com, foxsports.com, mutualofomaha.com and foxnews.com. Researcher Kirk (Ohey) reports 51 traced sites overall; only 15 captures have body-level verification. Findings from Kirk were not independently re-executed by Hudson Rock.
The main payload, track.js from 1688360.cc (live variant 29,882 bytes), performs three browser-side functions. (1) Cryptocurrency address substitution: replaces EVM, Bitcoin, TRON, TON, Solana, Litecoin, Bitcoin Cash and Dogecoin addresses in visible text, inputs, links, copy events and clipboard API calls using a 50 ms scan plus mutation observation; no seed extraction or automatic transfer logic was found. (2) Card skimming: hooks form submissions and fetch/XMLHttpRequest POST bodies, validates card prefix, Luhn, expiry and CVV, and sends n/e/c/u fields to 1688360.cc/card.php (sendBeacon when available); no actual card transmission was demonstrated and bypass of hosted-payment-frame isolation was not shown. (3) Conditional collection: posts page URL, document.cookie (JS-readable only, up to 8,000 chars), recorded address matches and an optional html2canvas viewport JPEG to 1688360.cc/hit.php. BBC and HSBC UAE assets embed a Base64 fallback copy (24,392 bytes) that runs if the remote track.js fails or has not completed after 2.5 seconds, so blocking the attacker domain alone does not neutralise the poisoned project code.
An iOS branch recognises iPhone/iPad/iPod user agents (and iPad-style Macintosh touch), injects two boot.js references with channel code 02111110, and for iOS 18.6 or lower adds JIUJIU888/weifile.html. Per Kirk's router logic, iOS 13-17 is routed to /A8/a.html (Coruna-associated), 18-25 to /D11/b.html (DarkSword-labelled), 26 to /C10/b.html (untested; C10 reported disabled earlier), above 26 no chain. URLScan captures with spoofed iOS 17.5 and 18.6 user agents received the A8 and D11 entry pages and packed JavaScript stages on eil786npto9u5bq.icu and 5vobauua3ojpklx.icu. JIUJIU888/index.js and the A8 script install globalThis.obChTK from a Base64 module map matching an existing Coruna detection rule; the D11 entry uses /api/gs_state and /api/pe_ok gates before loading frame.html, consistent with the public GTIG description of DarkSword delivery. Static markers include ds_vk, __DAILI_CFG, chainFrame with data-daili-gate=v1, XOR key d7K9mP2x, a 24-hour frame gate and a 30-minute pause after two failures. The infrastructure is linked to the Caishen iOS C2 V2 kit described in earlier Ohey research; this supports kit reuse, not operator identity.
Context from primary sources: Google GTIG (2026-03-03) documented Coruna as five full iOS chains / 23 exploits for iOS 13.0-17.2.1 used by a surveillance-vendor customer, UNC6353 (suspected Russian espionage) and UNC6691 (financially motivated, China-based), with PLASMAGRID wallet-theft payload. GTIG (2026-03-18) documented DarkSword (six vulnerabilities, iOS 18.4-18.7, patched by iOS 26.3) used by UNC6748, PARS Defense and UNC6353 with GHOSTKNIFE/GHOSTSABER/GHOSTBLADE payloads. Those kits are third-party reporting; no CVE is established as exploited in THIS campaign. Delivery, completed response, exploit execution and data theft are separate findings, and only delivery and code behaviour are established. The source recorded the poisoned BBC response disappearing by ~21:10 UTC on 2026-11-16 while GitLab, Zoom and minute-ly loaders were still present.
MITRE ATT&CK techniques by tactic:
Defense Evasion
T1027 Obfuscated Files or Information; T1140 Deobfuscate/Decode Files or Information; T1480 Execution Guardrails
Exfiltration
T1041 Exfiltration Over C2 Channel
Collection
T1056 Input Capture; T1113 Screen Capture; T1115 Clipboard Data; T1185 Browser Session Hijacking
Execution
T1059.007 JavaScript; T1203 Exploitation for Client Execution
Command and Control
Initial Access
T1189 Drive-by Compromise; T1190 Exploit Public-Facing Application; T1195.002 Compromise Software Supply Chain; T1659 Content Injection
Credential Access
T1539 Steal Web Session Cookie
Impact
T1565.002 Transmitted Data Manipulation; T1657 Financial Theft
Resource Development
T1583.001 Acquire Infrastructure; T1588.005 Obtain Capabilities; T1608.004 Stage Capabilities
Remediation
Patches
- Update iOS to the latest release; GTIG states Coruna does not work on current iOS and DarkSword is fully patched in iOS 26.3
Immediate actions
- Preserve response bytes, hashes, headers and timestamps of poisoned third-party scripts, then remove injected Optimizely projectJS code and review Kissmetrics, StorifyMe and minute-ly usage
- Block 1688360.cc, webshell.lol, 5vobauua3ojpklx.icu, eil786npto9u5bq.icu, je4ft3gn8vxudum.icu, ios-hack.icu and l7bcsnwc7nzr3uo.icu at DNS/proxy/egress
- Hunt for POSTs to /card.php and /hit.php and for boot.js, /A8/a.html, /D11/b.html, /api/gs_state and /api/pe_ok requests as separate events
- Verify what visitors actually receive from each third-party script (edge-cache state varies) and capture network traces, not only final DOM
Workarounds
- Temporarily disable affected third-party scripts on payment, login and account pages until verified clean
- Enable Lockdown Mode on iOS devices that cannot be updated
Longer-term hardening
- Apply Subresource Integrity where feasible and a strict Content-Security-Policy (script-src/connect-src allowlists, report-to) for payment, login and account pages
- Audit third-party script publishing paths: Optimizely project history, account activity, API credentials, integrations and SSO/MFA for vendor consoles
- Inventory and minimise third-party scripts on payment pages; use hosted payment frames with strict isolation
- Keep iOS devices updated (iOS 26.3 or later patches all DarkSword vulnerabilities) and use Lockdown Mode for high-risk users
Affected and fixed versions
- Optimizely — Customer projectJS (cdn.optimizely.com)
Vulnerable versions: Poisoned project files observed 2026-11-16 (BBC, Zoom, GitLab, HSBC UAE, Hootsuite, Virgin Media O2, Stamps.com, Tricentis, Salesforce, CNN, Mutual of Omaha, Fox News) - Kissmetrics — i.kissmetrics.com/i.js, i.kissmetrics.io/i.js, scripts.kissmetrics.io
Vulnerable versions: Poisoned responses observed 2026-11-13 - StorifyMe — storifyme-elements.min.js and custom widgets
Vulnerable versions: Poisoned responses observed 2026-11-13 to 2026-11-15 - minute-ly — apv-static.minute-ly.com loader.js
Vulnerable versions: Poisoned loader observed 2026-11-16 - Apple — iOS (targeted by routed exploit-kit entries)
Vulnerable versions: iOS 13-17 (Coruna-associated A8 route); iOS 18-25 (DarkSword-labelled D11 route)
Fixed in: iOS 26.3 or later (DarkSword per GTIG)
Timeline
Last updated · 2 updates · revalidated 2× · latest source
- Google GTIG publishes Coruna iOS exploit kit analysis (iOS 13.0-17.2.1; UNC6353 and UNC6691 use); context for the A8 route observed later.
- GTIG, iVerify and Lookout disclose the DarkSword iOS chain (iOS 18.4-18.7); fully patched in iOS 26.3.
- locked-ioc-32, the same actor's earlier injector, first appears on Turkish sites via an SEO 'backlink panel' script (locked-ioc-34). The TRON wallet has been in use since March 2026 (Report URI).
- locked-ioc-22 (iOS exploit router) registered at 17:52 UTC.
- Earliest verified poisoned Kissmetrics response captured on securitybank.com (05:42 UTC); BuzzStream, chicagoitsupport.com and a poisoned StorifyMe widget on faz.net follow the same day.
- Poisoned StorifyMe component captured on smh.com.au and unibas.ch; vidaxl.co.uk StorifyMe widget captured 2026-11-15.
- Optimizely states the actor exploited a vulnerability in the Experimentation account signup flow to add malicious code to customer projects; it blocks the vulnerable path, deploys a fix, removes injected code, ends sessions, revokes tokens, resets access for affected accounts, blocks delivery via Cloudflare and notifies customers. Report URI states this was not CDN cache poisoning.
- ~20:40-21:10 UTC: mass cleanup (Ledger 20:41, Smartsheet 20:56, BBC 21:07); by 21:52 UTC only 1 of 188 snippets (TalkTalk) still contained malicious code. Optimizely had not responded.
- 19:24-19:59 UTC: Report URI sweep finds 186 of 1,611 snippets infected (final tally 188 snippets across 178 organizations); about $7,040 had reached attacker wallets by 19:25 UTC.
- 16:38 UTC locked-ioc-23 registered; 17:13-17:57 UTC second iOS host added and Coruna iframe injected into every iOS 18.6-or-earlier visit.
- 13:03-13:05 UTC: attacker sweeps stolen crypto to exchanges (USDC to a HitBTC deposit address). Card skimmer (card.php) added to track.js at 13:43 UTC.
- 04:21 UTC: first Optimizely injection seen (BBC snippet loading locked-ioc-32); last clean capture 02:21 UTC. locked-ioc-36 registered 05:23 UTC; first track.js build 08:10 UTC.
- Hudson Rock publishes the technical brief via Infostealers.com.
- At ~21:10 UTC a BBC response was observed without the injection while GitLab, Zoom and minute-ly loaders remained; not a global remediation time.
- iOS 18.6 UA scan returns D11 entry and JIUJIU888 stage (21:08:43 UTC); iOS 17.5 UA scan returns A8 entry and packed Coruna-associated script (21:12:49).
- Hudson Rock retrieves live track.js (29,882 bytes) and poisoned GitLab, BBC and HSBC UAE assets (21:00-21:02 UTC), then five more Optimizely assets and the minute-ly loader.
- urlscan shows GitLab marketing page delivering track.js via Optimizely (16:17 UTC); HSBC UAE archive shows project file and track.js reference (17:16-17:29).
- Poisoned Optimizely projectJS captured on BBC (04:31 UTC), Zoom (11:41), Salesforce, CNN, Fox Sports (minute-ly), Mutual of Omaha and Fox News through the day.
Update history
- 2026-11-17 — Optimizely Supply Chain Attack Injects Malicious JavaScript (track.js) into Customer Sites Including CNN, Fox News, BBC and Salesforce: What changed No field escalation applied. Severity stays HIGH and exploitability ACTIVE. The newer report rates the incident CRITICAL, but that is its own rating and the evidence it adds is mostly root cause and remediation, not new impact.
- 2026-11-17 — Optimizely Web Supply Chain Attack Injects track.js Crypto-Clipper/Card Skimmer and Coruna/DarkSword iOS Exploit Kit Loader into 178 Organizations' Sites: What changed Scope widened from 21 named sites to 188 infected Optimizely snippets across 178 organizations (of 1,611 analyzed).
Sources cited
- Poisoned Scripts on BBC, Zoom and Other Major Websites Linked to iPhone Exploit Loaders (Hudson Rock / Infostealers.com) infostealers.com · Research · 2026-11-16
- Coruna: The Mysterious Journey of a Powerful iOS Exploit Kit (Google GTIG) Google Cloud · Research · 2026-03-03
- DarkSword iOS Exploit Chain (Google GTIG) Google Cloud · Research · 2026-03-18
- Coruna: Inside the Nation-State Grade iOS Exploit Kit (iVerify) iverify.io · Research · 2026-03-04
- A DarkSword hangs over unpatched iPhones (Malwarebytes) malwarebytes.com · Blog · 2026-03-19
- Coruna iOS Exploit Kit Uses 23 Exploits (The Hacker News) The Hacker News · News · 2026-03-04
- DarkSword: Researchers uncover another iOS exploit kit (Help Net Security) Help Net Security · News · 2026-03-19
Further reading
Drafted by the Threadlinqs research pipeline from 7 sources; published ; last updated .