
Hundreds of websites using the Optimizely website personalization platform were quietly serving visitors a malicious JavaScript package that could steal cryptocurrency, skim payment card data, capture cookies and screenshots, and even push iPhone and iPad exploit kits, researchers said Tuesday. The campaign affected 178 websites identified by ReportURI, including high-profile properties such as CNN, Zoom, the BBC, and Salesforce.
Optimizely’s infrastructure became the delivery channel
According to ReportURI, the malicious code was served through JavaScript digitally signed by Optimizely and then loaded into customer pages. That meant the tainted scripts were delivered from legitimate Optimizely infrastructure, including cdn.optimizely.com, rather than through the more familiar route of a simple compromised third-party ad or analytics tag.
Scott Helme of ReportURI said the mechanism was not a CDN cache poisoning attack, which would rely on HTTP responses being cached in a harmful form. Instead, he said the attack made use of Optimizely’s own infrastructure and domains to distribute the modified code. In an added complication, Optimizely’s own customer JavaScript for the www.optimizely.com and login.optimizely.com URLs was also affected.
Helme said the key mystery is how the attacker gained write access to so many unrelated customer projects. He wrote that possibilities include “a compromised Optimizely system, a compromised credential with cross-customer access, or something else,” and added, “That’s for Optimizely to explain.” Optimizely representatives did not respond to email questions, according to the report.
What the malicious scripts were doing
The scripts were named track.js, and ReportURI said all four malicious components became active at the same time. The payloads behaved differently depending on the victim’s platform and what was happening on the page where the code executed.
In practical terms, the attack combined several forms of theft in one package:
- Crypto wallet address swapping: The script searched for wallet addresses tied to nine different cryptocurrencies and replaced them with addresses controlled by the attackers. ReportURI said funds have already been stolen from some victims whose devices ran the code.
- Card skimming: The script watched HTTP requests for payment card numbers, expiry dates, and CVVs. When all three fields were present, it sent the data to
https://1688360.cc/card.php. Card entries made inside hosted payment iframes such as Stripe and Adyen were not affected. - Cookie and screenshot theft: When the script swapped an address next to wallet or payment wording, it sent the victim’s
document.cookie, the original address it replaced, and a screenshot of the page tohttps://1688360.cc/hit.php. - iPhone and iPad exploit delivery: On iOS devices, the code delivered exploit kits known as Coruna and DarkSword.
The breadth of the campaign makes it especially unusual. Rather than focusing on one kind of theft, the attackers appear to have assembled a multipurpose malware cocktail designed to extract value from different types of users and devices in the same browsing session.
Why the iPhone and iPad angle stands out
One of the more concerning parts of the attack was the iOS-specific component. ReportURI said the exploit kits were delivered through a file named boot.js served from the domains 5vobauua3ojpklx.icu and eil786npto9u5bq.icu. That file delivered a hidden iframe containing an exploit chain tailored to the iOS version being used.
According to the researchers, Coruna can infect iOS versions 13 through 17, while DarkSword targets iOS 18.x. A separate chain for iOS 26 fingerprints the device but does not appear to do anything else. The report also noted that earlier this month, researchers saw hackers possibly linked to the Kremlin attempting to deliver DarkSword through multiple vectors, including a compromised web analytics service.
That does not mean every victim device was successfully compromised. The report is clear that the available evidence does not yet show how successful the exploit chains were at installing payloads. Still, the presence of exploit kits in a campaign that already includes wallet theft, card skimming, and browser data exfiltration raises the stakes for anyone exposed to the tainted scripts.
What website operators and users should know
The list of affected sites, while already large, may not be complete. ReportURI said the 178-site tally could be an undercount, and Helme noted in an interview that “Ars wasn’t affected as far as we saw, and I haven’t seen any other indications of it.”
For site operators, the immediate lesson is that trusted platform code can become a high-leverage attack vector when attackers get a foothold in the delivery chain. Because Optimizely snippets run on every page the customer serves, any compromise there can affect a site far beyond a single page or subdomain.
Administrators looking for signs of compromise were pointed to ReportURI’s post and a separate note from security firm Cynet for indicators of compromise. Those indicators may help website owners or security teams identify whether their projects were touched by the malicious JavaScript.
For users, the implications are less straightforward. Visitors to affected pages could have been exposed without clicking anything unusual, and the code was capable of acting on the page in several ways at once. The indicators may help some people determine whether they were affected, but the technical skill required to interpret them is likely beyond many ordinary users.
What remains unanswered
As with many breaking reports about mass malware activity, there are still major gaps. ReportURI said it does not yet know how the attacker obtained write access across so many separate projects, whether the compromise involved Optimizely itself, a credential with broad privileges, or another route entirely. There is also no solid public information yet on the success rate of the exploit chains or the full scope of affected sites.
That uncertainty does not reduce the seriousness of the event. Even based on the limited facts available, the campaign shows how a single compromised distribution channel can be used to hit hundreds of websites at once while mixing financial theft, credential harvesting, and mobile exploitation attempts in one coordinated operation.
For now, the story is still unfolding. Researchers say more precise details may emerge as evidence is analyzed, but the current picture is already stark: a widely used website personalization platform became an unwitting conduit for malware that reached readers and shoppers on some of the internet’s most recognizable sites.
Source: Original report
Was this helpful?
Explore more: Application Audit & Review More Cybersecurity Tech News
Last Modified: November 18, 2026 at 10:31 pm
0 views
