Optimizely Supply Chain Attack
On 16 November 2026, an attacker exploited a vulnerability in Optimizely's Experimentation account signup flow and planted malicious JavaScript in customer projects. Optimizely's own CDN then served that code to visitors of 178 organisations' websites, including the BBC, Zoom, CNN, HSBC, Salesforce and IKEA. The code swapped crypto wallet addresses, skimmed payment cards, grabbed cookies and screenshots, and tried to push exploit kits onto older iPhones and iPads. Optimizely says the flaw is fixed and the code is gone.
None of those 178 organisations had to be hacked for this to happen. Their websites were simply the delivery route. That's what makes this one worth paying attention to.
Here's what we know so far.
What happened in the Optimizely attack?
Security researcher Scott Helme and his team at Report URI spotted it on 16 November. They checked 1,611 Optimizely scripts and found malicious code in 188 of them, tied to 178 organisations. The real total could be higher. They could only check the scripts they could find.
Some quick context. Optimizely is a platform companies use to run website experiments, personalise content and A/B test layouts. To make that work, you embed a small JavaScript snippet on your site. It loads from Optimizely's servers and runs in every visitor's browser.
Every Optimizely Web project has a "Project JavaScript" field. It's custom code that runs on every page where the snippet loads. That's where the attacker planted their loader, which then pulled further code from attacker-controlled infrastructure. Report URI is clear this wasn't cache poisoning. Optimizely's CDN served exactly what was in the projects. The projects themselves had been tampered with.
The BBC gives a good sense of the timeline. Report URI captured a clean BBC snippet at 02:21 UTC, an injected version at 04:21 UTC, and the main payload loading on bbc.com by 08:28 UTC. Over the course of the day, the attacker rebuilt the payload 26 times, adding features as they went.
Who was affected?
The list of affected URIs spans financial services, tech, retail, media and aviation. Named organisations include:
- BBC
- Zoom
- CNN
- HSBC (sites in 11 markets)
- Salesforce
- Equifax
- IKEA
- GitLab
- Air France-KLM
- Ledger
- Peacock
- Fox News
Even Optimizely's own websites were carrying the code.
Being on this list doesn't mean an organisation's internal network was breached. What's confirmed is that malicious scripts were delivered to visitors through sites running Optimizely.
Responses so far have varied. Salesforce disabled Optimizely scripts across its preview and production sites as a precaution. Equifax stopped using the platform and says it's found no evidence its systems were compromised. HSBC says unauthorised scripts weren't active on its websites.
What was the malicious code designed to do?
This wasn't vandalism. It was a multi-purpose kit built to make money, and it had four jobs.
Steal cryptocurrency
It swapped crypto wallet addresses that visitors copied for addresses the attacker controlled. Paste, send, gone. Report URI traced roughly $7,000 into the attacker's wallets by the evening of 16 November, so this is confirmed theft, not a theoretical risk.
Skim payment cards
It captured card details typed into forms on affected pages. Helme has been upfront that there's no way for his team to see how many cards were taken.
Grab cookies and screenshots
It collected any cookies JavaScript could read and took screenshots of the page. If a site kept session tokens in cookies that weren't set to HttpOnly, those sessions should be treated as exposed.
Attack iPhones and iPads
It also tried to deliver two iOS exploit kits, Coruna and DarkSword. Report URI maps Coruna to iOS 13 to 17 and DarkSword to iOS 18, so fully updated devices weren't the target. No successful device compromise has been demonstrated so far. But it shows how a trusted website can become a route into its visitors' devices, not just the business behind it.
How did Optimizely respond?
Optimizely confirmed that an unauthorised party exploited a vulnerability in its Experimentation account signup flow to add malicious JavaScript to some customer projects. It says it has:
- Blocked access to the vulnerable path and deployed a fix
- Removed the injected code from affected projects
- Ended active sessions, revoked tokens and reset access for affected user accounts
- Added protections to block unauthorised code delivery
- Notified affected customers
The clean-up was quick once it started. At 20:56 UTC on 16 November, 169 of the scripts Report URI was tracking were still infected. By 21:52 UTC that was down to one, and by late that evening all of them were clean. The attacker's main payload domain has since been suspended. The investigation is still ongoing.
What do we still not know?
How the attacker got in so widely. Optimizely has named the signup flow as the entry point. Report URI says it still doesn't know how the attacker gained write access to so many unrelated projects, or why those particular projects were picked.
How many people were exposed. We know 178 organisations. There's no figure yet for individual visitors.
How much data was stolen. The roughly $7,000 in crypto is the only hard number. Card details and cookies taken are unknown.
Whether any devices were compromised. The iOS exploits were attempted. Whether any landed hasn't been established.
What should organisations take away from this?
Third-party JavaScript gets a lot of trust. It runs with the same access as your own code, in your visitors' browsers, on your domain. If the supplier behind it is compromised, so is every site loading it. Here's where I'd focus.
1. Know your third-party suppliers
Make a list of every external platform that can put code on your website or touch your data. Who holds admin access to those accounts? What assurance do you have over the supplier's security? If you're working towards frameworks like the DSPT or Defence Cyber Certification, supplier management is already part of what you'll need to evidence.
2. Monitor the scripts running on your site
This attack was caught because someone was watching scripts change in real time. Most organisations would never have noticed. Know what's running, and get alerted when a script changes or starts talking to a domain you've never seen.
3. Get your Content Security Policy right
A properly enforced CSP limits which scripts can run and where data can be sent. But the detail matters. Report URI found that on one affected site, the CSP blocked the external payload and the data theft, yet an inline backup copy still ran because the policy allowed eval. That's a setting Optimizely's own documentation recommends. Restrict connect-src, avoid 'unsafe-eval' wherever you can, and turn on reporting so you actually see violations.
4. Test regularly
A penetration test won't stop a supplier getting breached. What it will do is check how your web apps, integrations and security headers hold up, and whether a rogue script could do as much damage as this one did. If you're in the defence supply chain, it's also a requirement.
5. Plan how you'd pull the plug
If a third-party script turned malicious at 4am, how fast could you switch it off? Who decides? Who tells customers? Build that into your incident response and business continuity planning before you need it.
If you use Optimizely, check these now
- Audit the Project JavaScript and any custom code in your experiments for anything you didn't write
- Rotate your Optimizely credentials and API tokens
- Rotate any session tokens or secrets stored in cookies that aren't HttpOnly
- Check whether Optimizely has contacted you as an affected customer
- Review your CSP, especially whether it allows 'unsafe-eval'
Your perimeter is bigger than you think
We keep seeing the same pattern: one trusted supplier, hundreds of victims. Strong internal controls matter. But they don't cover the code a supplier pushes into your visitors' browsers.
So the question isn't just "are our systems secure?" It's "is everything we let run on our site secure too?"
We'll keep following this and update the post as more comes out.
This article reflects information available as of 17 November 2026. Details may change as Optimizely and security researchers publish further findings.
FAQs
What is the Optimizely supply chain attack?
On 16 November 2026, an attacker exploited a vulnerability in Optimizely's Experimentation account signup flow and planted malicious JavaScript in customer projects. Optimizely's CDN then served that code to visitors of 178 organisations' websites. It was found by Scott Helme and his team at Report URI.
Which websites were affected by the Optimizely attack?
Report URI found malicious code in 188 of the 1,611 Optimizely scripts it checked, linked to 178 organisations. Named sites include the BBC, Zoom, CNN, HSBC, Salesforce, Equifax, IKEA, GitLab, Air France-KLM, Ledger, Peacock and Fox News. Being on the list doesn't mean an organisation's own systems were breached.
I visited an affected website. Was my data stolen?
Not necessarily. Not every visitor will have run the malicious code, and nobody yet knows how many card details or cookies were taken. If you visited an affected site on 16 November, update your iPhone or iPad, double-check any crypto transfers you made, and keep an eye on your card statements.
Has Optimizely fixed the vulnerability?
Optimizely says it has blocked the vulnerable path, deployed a fix, removed the injected code, ended active sessions, revoked tokens, reset access for affected accounts and notified affected customers. Report URI says the infected files it was tracking were clean by late on 16 November.
How can organisations protect themselves against malicious third-party JavaScript?
Know exactly which third-party scripts run on your site and who controls them. Monitor for unexpected changes, enforce a Content Security Policy that restricts where data can be sent and avoids 'unsafe-eval' where possible, test your web apps regularly, and have a plan to switch off a compromised integration fast.
Related reading
- Does DSPT Require Penetration Testing?
- DCC Penetration Testing Requirements: A Complete Guide to Control 2403
- Binance's $6.3 Billion Security Incident: What We Know, and What May Have Happened
- Google Cloud Outage: What It Means for Your Business Continuity Plan