On 16th November 2026, Optimizely Web Experimentation served visitors to websites belonging to 178 organisations malicious JavaScript.
The affected organisations had not necessarily suffered a breach of their own infrastructure. Instead, the malicious code arrived through a third-party service they had already approved and trusted. Research by Scott Helme at Report URI identified 188 infected snippets among the 1,611 Optimizely snippets examined.
The malicious code could replace cryptocurrency payment addresses, steal card details entered into payment forms, capture cookies and screenshots, and deliver exploit kits targeting older iPhones and iPads; the exposure lasted more than 13 hours.
Optimizely confirmed that an attacker had exploited a vulnerability in its account signup flow to add JavaScript to customer projects. It subsequently blocked the vulnerability and removed the injected code.
The incident highlights an uncomfortable reality: trusting a third-party supplier doesn’t mean everything it delivers will remain trustworthy.
Every affected website continued doing what it had always done: loading a script from Optimizely.
The script came from the legitimate domain, over a valid certificate. The malicious code had been added to a field customers legitimately use to run their own JavaScript across their websites; the browser therefore treated it as trusted.
This is why traditional security controls struggled.
Report URI found that only around 15% of the affected hosts it checked had a Content Security Policy (CSP) that stopped the payload. Allowing the Optimizely domain, which these websites needed to do to use the service, offered no protection when the malicious code was delivered through that same domain.
The lesson isn’t to stop using third-party services; modern websites depend on them, it’s to recognise that trusting a supplier is not a one-off decision. What sits behind that relationship can change at any time.
Trust Monitor maps the third-party services your live websites load and connect to, listing each one in your supplier risks view with an associated risk level.
A script running across every page, with access to that page, represents a high-trust relationship. Trust Monitor assigns these relationships a trust level and shows which of your websites load the service. This gives security teams visibility into questions such as: should an experimentation script have access to our login and payment pages?
It’s a question they can investigate and act on before an incident occurs.
On 16 November, the supplier hadn’t changed. What sat behind it had.
Websites that had previously loaded Optimizely began pulling a script from a new domain, registered that same morning, and sending data to it. Trust Monitor doesn’t stop at the first scan; it continues checking what your websites load and connect to.
When a new, previously unseen domain appears behind a supplier you’re already monitoring, Trust Monitor can detect and alert you to that change, even if the original supplier domain remains the same. So did the script address and certificate.
The important signal was the new relationship behind it. That’s what Trust Monitor is designed to identify.
When an incident like this happens, one of the first questions security teams need to answer is: are we affected?
In Trust Monitor, search for the Optimizely domain to identify every website in your estate that loads it. This gives you a clear starting point for investigating potential exposure, rather than relying on manual checks and emails between web teams.
There is one important caveat: Some organisations serve their Optimizely snippets from their own domains. These websites won’t appear in a search for the Optimizely domain, so they need to be checked separately.
If you use Optimizely Web, you should also:
Scott Helme’s Report URI investigation includes the full response checklist and the attacker domains to look for.
An allow-list tells your browser which domains you trust. It can’t tell you when that trust is no longer deserved. The organisations caught up in this incident were using a legitimate third-party service as intended. That’s what makes the incident such an important lesson. The same type of attack could happen through any script, tag or widget your websites depend on.
You can’t remove all third-party risk. But you can improve your visibility of it. Know which suppliers your live websites rely on. Understand the risk associated with those relationships. And, crucially, know when something changes.
Find out how Trust Monitor can help you uncover your third-party dependencies and monitor what changes.
REQUEST YOUR FREE TRUST MONITOR SCAN
The research into this incident was carried out by Scott Helme and the team at Report URI.
No software to download or install.
Contact us or call us 0113 887 8380
AppCheck is a software security vendor based in the UK, offering a leading security scanning platform that automates the discovery of security flaws within organisations websites, applications, network and cloud infrastructure. AppCheck are authorised by the Common Vulnerabilities and Exposures (CVE) Program as a CVE Numbering Authority (CNA)