Inside the Optimizely Supply-Chain Hijack: How One Compromised CDN Poisoned 178 Global Brands
When a user visits a major news outlet, an international bank, or an enterprise SaaS portal, their browser silently executes dozens of third-party scripts before the page finishes rendering. On November 16, 2026, threat actors turned that implicit architectural trust into a global malware distribution network.
By compromising customer projects on Optimizely Web Experimentation—a ubiquitous A/B testing platform deployed across roughly 190,000 websites—attackers injected malicious JavaScript loaders directly into Project JavaScript snippets. Because those files were served directly from cdn.optimizely.com and signed with Optimizely’s legitimate TLS certificate, the poisoned code bypassed conventional domain-reputation filters and landed straight inside the browsers of millions of unsuspecting visitors.
Security researcher Scott Helme of Report URI analyzed 1,611 Optimizely snippets during the incident window and uncovered 188 infected instances spanning at least 178 organizations. Banking and financial institutions made up the single largest target vertical with 32 affected organizations, alongside household names across media, retail, and technology—including CNN, the BBC, Fox News, HSBC, IKEA, Zoom, Salesforce, Equifax, GitLab, Air France-KLM, and Ledger. Even Optimizely’s own public website and customer login portal briefly served the malicious loader.
Once executed in a visitor’s browser, the second-stage payload functioned as a multi-purpose Swiss Army knife designed to monetize both desktop and mobile traffic simultaneously:
-
Checkout Form Skimming: The script monitored payment forms on affected e-commerce and service portals, silently exfiltrating credit card numbers and billing details as users typed them.
-
Session & Screenshot Theft: When the script detected payment or wallet-related keywords on a page, it harvested active session cookies and captured browser screenshots to enable downstream account takeover.
-
Cryptocurrency Address Swapping: The malware scanned web pages for cryptocurrency wallet addresses and dynamically replaced them with attacker-controlled destinations, netting roughly $7,040 in tracked inflows within hours of deployment.
-
Mobile Exploit Delivery: Rather than ignoring mobile visitors, the payload profiled incoming user agents and served two distinct iOS exploit kits—dubbed Coruna and DarkSword—to compromise visiting iPhones and iPads, triggering repeated 60-second page reloads on devices running iOS 18.6 or earlier.
Optimizely purged the malicious snippets and closed the exploited path by late November 16, roughly 18 to 20 hours after researchers first spotted the tampered code on the BBC’s website. Yet the incident exposes a glaring blind spot in modern web security: runtime third-party script execution.
For Purple Teams, the Optimizely hijack proves that perimeter firewalls and secure internal CI/CD pipelines offer zero protection when an approved marketing or analytics vendor becomes the delivery vehicle. Static Subresource Integrity (SRI) hashes—the standard defense against CDN tampering—rarely work for A/B testing tools because the vendor dynamically updates the script whenever a marketing team tweaks an experiment.
Defending against this class of client-side supply-chain attack requires isolating sensitive workflows from third-party code entirely. Organizations must enforce strict Content Security Policies (CSP) to restrict unauthorized outbound data exfiltration, strip third-party experimentation tags from authentication and payment-processing pages, and deploy client-side behavioral script monitoring to detect unauthorized DOM manipulation in real time.
