
Security researcher Scott Helme of Report URI found a malicious script injected into 178 organisations through a popular A/B testing platform, according to the report. Helme audited 1,611 scripts and discovered the compromised code in 188 snippets across the affected companies. The list of victims includes major names like CNN, HSBC, the BBC, Zoom, and Salesforce.
The Attack Exploited A Trusted Domain
The malicious code arrived via Optimizely, a platform used by the victims to run on-page experiments. Attackers injected the script into customer project settings, which allowed the code to run on every site that loaded Optimizely’s library. Because the script came from a trusted vendor domain, browsers and security defenses accepted it without raising an alarm. The vendor confirmed that an unauthorised party exploited a bug in its account sign-up process to gain access to these settings.
The vendor has since patched the sign-up bug and revoked tokens for affected accounts. Despite this cleanup, even the vendor’s own websites were found serving the loader. This situation highlights a fundamental risk in modern web development: when a company relies on third-party scripts, it effectively hands a master key to every room on its website and hopes nothing goes missing.
Read Also: Kresus Appoints Wall Street Veteran as CRO
The attacker used the compromised script to steal sensitive information. Helme’s analysis shows the code swapped copied cryptocurrency wallet addresses for ones the attacker controlled. It also scraped credit card details from form fields on the page. On iPhones and iPads, the script pulled in additional files that Helme says deliver exploit kits, though those files were not analyzed. The card skimmer appears unable to read payment details typed into hosted payment frames, which are isolated from the page’s own JavaScript environment.
The injection began around 04:21 UTC on 16 November, and most sites were exposed for approximately 16 hours. This timing means that for most of a standard business day, visitors to these sites were unknowingly submitting data to a compromised source. Traditional security systems often monitor a company’s internal network but cannot easily inspect traffic that originates from a trusted third-party domain, which explains why internal defenses failed to detect the breach.
A Vulnerable Supply Chain
Every organisation on the list had its own monitoring systems in place, yet they all missed the compromise. The script arrived from a source with a recognizable vendor name, so nothing in the logs looked suspicious. This incident illustrates the difficulty of securing a supply chain. When a website relies on scripts written by another company, that business effectively leaves the door unlocked for anyone who can compromise the vendor.
