Optimizely Web malware: Report URI finds 178 organisations served crypto and card stealer
Report URI says malicious JavaScript was served through Optimizely Web snippets on 16 November to sites including Zoom, the BBC and CNN. Optimizely has not commented; here is what is claimed and what to check.
· 6 min read
Written by AI from the sources linked below. How we use AI
On this page
Security researcher Scott Helme of Report URI says that on 16 November 2026 malicious JavaScript was served from Optimizely's own CDN to visitors of 178 organisations, among them Zoom, the BBC, CNN, Salesforce, GitLab, IKEA, HSBC and Equifax. According to his write-up, the code swapped cryptocurrency wallet addresses, skimmed card details typed into site forms and, on older iPhones and iPads, loaded iOS exploit code. As of publication Optimizely had not answered Report URI, and we found no statement from the company. Almost everything below therefore rests on one researcher's account, which we label as such throughout.
Key points
- Report URI says it checked 1,611 Optimizely Web snippets and found 188 infected, belonging to 178 organisations; it expects the real total to be higher.
- The injected code sat in the "Project JavaScript" field of customer projects and was served from
cdn.optimizely.comwith valid certificates, so Report URI says this was not CDN cache poisoning. How the attacker got write access is unknown. - The payload swapped crypto wallet addresses, captured card number, expiry and CVV typed into a site's own forms, and sent cookies and screenshots to a server at
1688360.cc. - Report URI counted about $7,040 in cryptocurrency received by the attacker's wallets. The number of stolen cards is unknown.
- By 21:52 UTC on 16 November only one of 188 infected snippets (TalkTalk's) was still infected, per Report URI.
What to do
If you run Optimizely Web: check every Project JavaScript field and experiment custom code for 1688360.cc, atob( or scripts you don't recognise. Audit the users, API tokens and integrations with write access and rotate their credentials. Report URI advises treating card data typed into your own forms on 16 November as potentially exposed, rotating any secrets or session tokens kept in cookies that scripts can read (non-HttpOnly), and considering whether to tell customers.
If you visited an affected site on 16 November: Report URI recommends updating iPhones and iPads to iOS 18.7.3 or later, or iOS 26.3 or later, which it says are patched against the exploit chains involved. Check any crypto address before sending money, and watch card statements if you paid by card on an affected site.
Block the infrastructure: Report URI lists 1688360.cc, webshell.lol, linkforge.cc and several .icu domains as indicators; the full list, the IP addresses, file hashes and wallet addresses are in its post.
What happened
According to Report URI's timeline (all times UTC, 16 November unless stated), the first Optimizely injection it saw was in the BBC's snippet at 04:21. The main malicious script, track.js, was first seen on bbc.com at 08:28. Zoom's snippet was modified at 11:20. At 13:03 the attacker swept stolen crypto to exchanges, and at 13:43 a card skimmer was added. Between 17:13 and 17:57 an iOS exploit iframe was added for visitors on iOS 18.6 and earlier.
Report URI's own sweep between 19:24 and 19:59 found 186 of 1,611 snippets infected, rising to 188. A mass clean-up followed between roughly 20:40 and 21:10, with Ledger, Smartsheet and the BBC among those cleaned. At 21:52 only TalkTalk's snippet remained infected. A secondary write-up says Smartsheet's snippet was cleaned and then reinfected with a better-hidden loader, which would suggest continuing access; we could not confirm that in Report URI's post. Optimizely's own sites, optimizely.com and login.optimizely.com, were also affected, Report URI says.
Report URI says it tried to contact Optimizely through several channels without a reply when it published. The post says how the attacker gained access is "for Optimizely to explain", and Report URI does not claim to know whether it was a compromise of Optimizely's systems, a stolen credential or something else.
Background
Optimizely Web is an experimentation product: customers add a script from Optimizely's CDN to their pages, and and, per Report URI, customer projects have a "Project JavaScript" field that the attacker wrote into. Any third-party script runs with the full privileges of the page that loads it, here on the pages of banks, airlines and newsrooms, so whoever can change the script can run code in every visitor's browser. That is why security teams ask for a Content Security Policy.
Report URI says the same actor was seen earlier. Its injector domain, webshell.lol, first appeared on Turkish sites on 17 September through an SEO "backlink panel" script. Between 13 and 16 November an earlier wave injected webshell.lol/get.js through Kissmetrics, StorifyMe and Minute Media scripts; affected sites in that wave included securitybank.com, faz.net, smh.com.au, vidaxl and foxsports.com, according to the post. Our recent coverage of other client-side and supply-chain incidents includes the trojanized Claude Code releases served by the Cursor marketplace.
How it works
Report URI describes three loader forms: a plain script tag, a base64-encoded (atob()) URL, and a full inline copy that runs through eval if the domain is blocked. The track.js payload was rebuilt 26 times on the day, so file hashes change quickly; the post lists two SHA-256 values for the 17:13 and 17:57 builds.
The payload did four things, according to the post:
- Crypto address swapping: it replaced valid wallet addresses (Bitcoin, Ethereum, TRON, Solana and others) in page content, forms, the clipboard and outgoing requests with the attacker's.
- Card skimming: from 13:43 it captured card number, expiry and CVV from a site's own forms and sent them to
1688360.cc/card.php. Cards entered in hosted payment iframes could not be reached. - Data theft: it sent the page URL, non-HttpOnly cookies, the replaced addresses and an html2canvas screenshot to
1688360.cc/hit.php. - iOS exploitation: on iPhones and iPads it loaded exploit chains named Coruna (iOS 13 to 17) and DarkSword (iOS 18.x). The post says iOS 18.7.3 and 26.3 and later are patched. A secondary account says Helme did not analyse the exploits themselves, so treat the iOS part as less well established than the rest.
The numbers
Report URI checked the Content Security Policies of 256 affected organisations' hosts. Only 24 of them, about 15%, fully blocked the payload. The post says that allowing 'strict-dynamic' or 'unsafe-eval' let the attack through, and that restricting connect-src stopped data theft but not address swapping, because that part runs entirely in the page.
type: bar
title: Optimizely incident, key counts reported by Report URI
unit: count
source: Report URI, 16 November 2026
Snippets checked | 1611
Snippets infected | 188
Organisations with infected snippets | 178
Affected hosts whose CSP checked | 256
Hosts whose CSP fully blocked the payload | 24
What it means
For site owners, the lesson is uncomfortable: a valid certificate, the right domain and a well-known vendor did not make the script safe. A strict CSP is the main defence Report URI points to, and fewer than one in six of the sites it checked had one that worked here. Limiting which hosts a page may send data to (connect-src) reduces theft, but a clipper that rewrites addresses in the page needs the script itself to be stopped, which means tight script-src rules, no 'unsafe-eval', and reporting so you notice when something new appears.
For Optimizely customers it is also a question about access control on a field that can run code on production pages: who can edit it, whether it requires review, and whether changes alert anyone.
For the vendor, the open questions are the ones customers will ask first: how write access was obtained across unrelated projects, what was changed, and whether it is closed.
What we don't know yet
- How the attacker got in. Report URI says it does not know; Optimizely has said nothing we could find.
- Independent confirmation. Aside from Report URI, we only found write-ups that repeat its findings. One of them says StepSecurity separately confirmed the chain on a GitLab page; we could not find that confirmation ourselves.
- The real scale. Report URI says its count of 178 organisations is likely low, and counts cited elsewhere (186, 188) refer to snippets or to moments in the clean-up rather than disagreeing outright.
- Victims. The number of stolen cards and cookies is unknown; the $7,040 figure covers only crypto received by the wallets Report URI monitored.
What to watch next
- A statement or incident report from Optimizely, and whether it confirms Report URI's timeline.
- Notices from affected companies, especially anyone handling card payments on 16 November.
- Apple and exploit-kit research on the Coruna and DarkSword chains.
- Whether the infected-snippet count drops to zero or reinfections appear.
Sources
- World's Largest Brands Infected With Malware Via Optimizely, Report URI (Scott Helme), 16 November 2026 (primary source)
- Malware Spread Through Optimizely Web Reached 178 Organizations, Report URI Says, LavX News, 16 November 2026
- Optimizely Malware Hit 178 Sites: What It Stole and What to Do, ShiftDelete.Net
Written by
The official account of the people who run audiotale.net: announcements, conversation starters and the newsroom's drafts. Posts written with AI say so. Every other account here is a real person.
Read next
Cursor marketplace served trojanized Claude Code 2.1.312 and 2.1.313: Fireworks key led to SSH backdoor, up to 5,000 users
Cursor says attackers used a key exposed by Fireworks AI to plant tampered Claude Code extension builds for about 11 hours on 29 October. Here is who is affected and how to check.
audiotale.net team7 min read
wfx-loot: 4,837 hijacked GitHub accounts probed 61,764 repos and stole CI secrets
Aikido says stolen GitHub tokens planted a credential stealer in thousands of repositories from 10 November; here is how to check yours and what to rotate.
audiotale.net team8 min read
cPGuard 5.88.12 to 5.88.19 update channel hijacked: backdoor and root secrets stolen, update to 5.88.25
Attackers pushed a malicious script through cPGuard's auto-update on 2 November; here is how to check your servers, update and rotate root credentials.
audiotale.net team5 min read