High severity

Optimizely flaw used to push crypto-stealing and card-skimming code to 178 organisations

Researchers say code planted through Optimizely's experimentation platform swapped crypto wallet addresses, skimmed cards and pointed iPhones at exploit kits, all delivered from a trusted domain.

By Mr Elite Published 8 min read

Severity
High
Status
Reported by outlets
Sources
3
Last checked

The short version

  • Optimizely confirmed on 17 November that an unauthorised party exploited a vulnerability in Experimentation to add malicious JavaScript to some customer projects.
  • Report URI found malware in 188 of 1,611 Optimizely snippets it checked, belonging to 178 organisations, and says the true number is likely higher.
  • The code swapped crypto wallet addresses, skimmed card details, stole cookies and screenshots, and delivered the Coruna and DarkSword iOS exploit kits to iPhones and iPads.
  • Report URI said only 24 affected organisations, or 15%, had a Content Security Policy strict enough to block the attack, according to Help Net Security.
  • Hudson Rock said successful exploitation of iPhones has not been demonstrated.

What happened

An attacker exploited a vulnerability in Optimizely's Experimentation platform on 16 November 2026 to plant malicious JavaScript that reached visitors of 178 organisations' websites, according to researchers at Report URI.

Optimizely, a digital experience platform, confirmed the incident in a status page update on 17 November. The company said an unauthorised party had exploited a vulnerability in Experimentation and used it to add malicious JavaScript to some customer projects. It said it had fixed the flaw, removed the injected code and taken steps to secure affected user accounts. SecurityWeek reported that Optimizely has not said how many customers were affected.

Scott Helme of Report URI uncovered the attack. His team checked 1,611 Optimizely snippets and found malware in 188 of them. Names on the list include Zoom, the BBC, CNN, Salesforce, GitLab, IKEA, HSBC, Equifax, Air France-KLM, Ledger and Peacock. Help Net Security reported that banks and financial firms were the largest group, with 32 organisations.

Report URI's timeline puts the first injection at 04:21 UTC on 16 November. Infosecurity Magazine reported that all infected files had been cleaned by 23:11 UTC. Helme also said the payload was rebuilt 26 times during the day as the attacker added features and wallets, Help Net Security reported.

Salesforce told Reuters it was investigating unauthorised Optimizely changes and disabling scripts as a precaution, according to Help Net Security. The attacker's main payload domain has since been suspended, Helme said.

Report URI said the real number of victims is probably higher, because it could only check the snippets it found in time.

Why it happened

Optimizely has said only that the weakness was a vulnerability in Experimentation. It has not published technical details of the flaw, how the attacker used it, or how the affected accounts were reached. Its statement that it secured affected user accounts suggests account-level access was involved, but the company has not spelled that out.

Report URI described what the attacker did with that access. The code went into the Project JavaScript field, an optional box where customers add their own script, which runs on every page. The rest of each snippet was left alone. The tampered files were still served from Optimizely's content delivery network (a network of servers that delivers site files quickly) with its certificate, so visitors' browsers treated them as trusted.

Separately, Hudson Rock and a researcher known as Kirk of Ohey traced poisoned scripts to three other services: Kissmetrics, StorifyMe and minute-ly, with the earliest tampered responses on 13 November. SecurityWeek reported it is unclear how the code reached those providers. The researchers said shared hosting and loader code does not prove one operator or one method of entry.

How this kind of failure works

Most modern websites load scripts from other companies: analytics, testing tools, chat widgets. Your browser runs those scripts with the same power as the site's own code. If a third party is compromised, every site that loads its script inherits the problem without any breach of its own systems.

Think of a shop that lets a delivery firm stock its shelves overnight. The shop locks its doors and checks its own staff, but if someone swaps goods inside the delivery firm's depot, the shop sells the tampered products under its own name. Customers have no way to tell.

Optimizely's design makes this easy to picture. Customers paste code into a project, and Optimizely serves it to every visitor. Anyone who can write to that field can run code on every page that loads the snippet. Access to the editing side of a platform like this is therefore as valuable as access to the website itself.

A Content Security Policy, or CSP, is a set of browser rules a site publishes to say which scripts may run and where data may be sent. It limits damage when a trusted source turns bad. It only helps if it is strict, and allowing a vendor's whole domain gives no protection when the poisoned file comes from that same domain.

Who is affected

The direct victims are the 178 organisations Report URI named, and probably others it did not find in time. Their visitors on 16 November are the people at risk, particularly anyone who typed card details into a site's own payment form. Cards entered in separately hosted payment fields were out of the skimmer's reach, according to Report URI.

Helme said there is no way to see how many cards were skimmed. Hudson Rock said it has not linked the crypto transfers to visitors of the affected sites. Report URI said the attacker's wallets received about $7,040 by 19:25 UTC on 16 November, while Help Net Security put the figure at about $7,150 before the wallets were emptied early on 17 November.

On iPhones and iPads, Report URI said the code delivered the Coruna and DarkSword exploit kits, and advised updating to iOS 18.7.3 or later, or iOS 26.3 or later. Help Net Security reported that devices on iOS 18.6 or older saw pages reload every 60 seconds. Hudson Rock said it has not reproduced the later iOS stages or shown that any real iPhone was compromised.

If you run a site that uses Optimizely, check your Project JavaScript for code you did not write and review account activity. Visitors who used card forms on affected sites on 16 November may wish to watch statements and contact their bank.

How to defend against it

Tick each step as you check it. Your progress is saved on this device.

What to learn from this

The concept: Third-party script trust (web supply chain risk)

When your page loads a script from another company, that script runs with full access to the page. It can read form fields, cookies and what is on screen. Your security is then partly decided by that company's security.

The trust is usually based on where the file comes from. That is why this attack worked: the malicious files were served from Optimizely's own domain and certificate. Checks that ask only whether a file comes from an approved address will pass a poisoned file.

The same pattern appears in other third-party code, such as analytics scripts, chat widgets and open-source libraries. Hudson Rock and Ohey found tampered scripts from three other services in the same period.

Trusting a vendor's domain is not the same as trusting every file that domain serves, so limit what third-party code can do and watch what it contains.

Quick check

Sign in to earn 10 XP for a right answer

Why did allowing Optimizely's domain in a Content Security Policy fail to stop this attack?

Timeline

  1. September 2026 The same injection code first appeared on Turkish websites, Help Net Security reported.
  2. 13 November 2026 Earliest tampered responses recorded from Kissmetrics, StorifyMe and minute-ly scripts, per Hudson Rock and Ohey.
  3. 16 November, 04:21 UTC First Optimizely injection seen in Report URI's timeline.
  4. 16 November, afternoon The card skimmer arrived, Help Net Security reported.
  5. 16 November, 23:11 UTC Report URI said all infected files had been cleaned.
  6. 17 November, early Optimizely posted a status update confirming the incident; the attacker's wallets were emptied.

What is not known yet

  • How the attacker exploited the vulnerability, and whether stolen or abused accounts were involved, has not been disclosed by Optimizely.
  • The total number of affected customers is not known; Optimizely has not given a figure and Report URI says its count is likely low.
  • How many payment cards were skimmed is not known.
  • Whether any iPhone or iPad was actually compromised has not been shown, according to Hudson Rock.
  • How code reached the Kissmetrics, StorifyMe and minute-ly scripts, and whether one operator is behind all of it, is unclear.

Questions people ask

What happened in the Optimizely attack?

Optimizely said an unauthorised party exploited a vulnerability in its Experimentation product and added malicious JavaScript to some customer projects. Report URI found the code on 178 organisations' sites on 16 November. The code swapped crypto wallet addresses, skimmed cards, and targeted iPhones with exploit kits.

Was the Optimizely vulnerability fixed?

Optimizely said it fixed the vulnerability, removed the injected code from affected projects and added protections against unauthorised code being delivered. Report URI said all infected files were clean by 23:11 UTC on 16 November. Technical details of the flaw have not been published.

Are iPhone users at risk from this attack?

Report URI said the code delivered the Coruna and DarkSword iOS exploit kits to iPhones and iPads, and advised updating to iOS 18.7.3 or later, or iOS 26.3 or later. Hudson Rock said it has not shown successful exploitation of a real device. Updating is still the sensible step.

How can a Content Security Policy help against this kind of attack?

A strict policy can stop injected code from loading extra scripts or sending stolen data to an attacker's server. Report URI found only 24 of the affected organisations had a policy strong enough to block this attack. Allowing the vendor's whole domain gave no protection.

Sources

  • Infosecurity Magazine: Optimizely Flaw Exploited to Serve Malicious Code to Visitors of 178 Organizations
  • Help Net Security: Malware spread through Optimizely hits 178 organizations, including BBC, CNN and Zoom
  • SecurityWeek: Optimizely Vulnerability Exploited to Serve Malware on Major Websites

Know someone who should see this?

Keep reading