Bay Street Wire
Tech & BusinessOpinion

The Optimizely Breach: A Masterclass in Supply Chain Fragility

Portrait of Naomi Frost
Naomi Frostcybersecurity & privacyNov 17AI
The Optimizely Breach: A Masterclass in Supply Chain Fragility

AI-generated image · Bay Street Wire

When a trusted experimentation tool becomes a malware delivery system, the security of the world's largest brands is compromised—proving that your stack is only as strong as its most obscure dependency.

In the modern enterprise, security is often treated as a perimeter problem. We harden our own servers, we audit our own code, and we vet our primary vendors. But as Report URI first reported, the November 16, 2026, breach of Optimizely Web demonstrates a more terrifying reality: your infrastructure can be perfectly secure, and you can still be a vector for malware because you blindly trusted a third-party JavaScript snippet.

According to reporting from Report URI, malicious JavaScript was served to visitors of 178 different organizations through the Optimizely Web experimentation platform. The breach reached a massive scale, affecting high-traffic global sites such as Zoom, the BBC, CNN, Salesforce, GitLab, IKEA, HSBC, Equifax, Air France-KLM, Ledger, and Peacock. Even Optimizely's own domains—www.optimizely.com and login.optimizely.com—were infected.

**The Mechanism of Trust**

To understand how this happened, one must understand how Optimizely Web functions. Customers load a JavaScript "snippet" from Optimizely's CDN (e.g., cdn.optimizely.com). This snippet contains the platform's core code and a specific field called "Project JavaScript" (projectJS), which allows customers to write custom code in their dashboard that executes on every page.

Report URI explains that the attacker did not breach the CDN itself—this was not a cache-poisoning attack. Instead, the attacker gained write access to the Project JavaScript field of numerous unrelated customer projects. Because the malicious code was embedded within the legitimate snippet, it was served from a trusted Optimizely domain with a valid Optimizely certificate. For organizations that served the snippet from their own domains—such as www.vodafone.com.au or opt.scheels.com—the result was the same: the content originated from the compromised Optimizely project.

**The Payload: A Multi-Purpose Malware Kit**

This was not a simple redirect or a nuisance pop-up. The attacker deployed a sophisticated, evolving toolkit delivered via a script called `track.js`. Report URI notes that the payload was rebuilt 26 times throughout the day, with the attacker continuously adding features and wallets.

Depending on the user's device and behavior, the malware executed several distinct attacks:

1. **Cryptocurrency Theft:** The script scanned for valid wallet addresses across multiple chains, including BTC, ETH, EVM, TRON, Solana, TON, Litecoin, Bitcoin Cash, and Dogecoin. It then swapped these addresses for the attacker's in page text, form fields, links, images, and the clipboard. It even intercepted outgoing fetch/XHR requests and the iOS share sheet. Report URI noted that attackers have already successfully stolen cryptocurrency, with verified transactions. 2. **Credential and Data Theft:** The kit was designed to skim payment cards and steal cookies and screenshots. 3. **iOS Exploitation:** For users on iPhones and iPads (specifically those on iOS 18.6 or earlier), the malware delivered the DarkSword and Coruna iOS exploit kits via injected iframes.

**The Timeline of an Attack**

The precision of the attack suggests a calculated rollout. Report URI's timeline shows the actor's injector, `webshell.lol`, first appeared on Turkish sites in September 2026. By November 13, an earlier wave of infections hit sites like foxsports.com, smh.com.au, and securitybank.com through other scripts (Kissmetrics, StorifyMe, and Minute Media).

On November 16, the Optimizely campaign hit full throttle: * **04:21 UTC:** The first Optimizely injection was detected on the BBC snippet. * **08:10 UTC:** The first build of `track.js` (the crypto clipper) was created. * **11:20 UTC:** Zoom's snippet was modified. * **13:03 UTC:** The attacker began sweeping stolen cryptocurrency to exchanges. * **13:43 UTC:** A card skimmer was added to the payload. * **17:13–17:57 UTC:** The Coruna iframe was injected for iOS users.

By 19:24 UTC, a sweep identified 186 of 1,611 Optimizely snippets as infected. While a mass cleanup began around 20:40 UTC—cleaning sites like Ledger and the BBC—Report URI noted that Smartsheet was cleaned and then re-infected, suggesting the attacker maintained persistent access during the remediation process.

**The Defender's Dilemma**

(Opinion) This breach is a textbook example of why the "trusted vendor" model is a security fallacy. We have built a web where a single line of code in a third-party dashboard can bypass the security posture of a Fortune 500 company. When you import a script from a CDN, you aren't just importing a feature; you are importing the entire security debt of that vendor and every person with access to their dashboard.

Report URI argues that this activity could have been detected through JavaScript integrity monitoring, which identifies unexpected changes to trusted scripts and suspicious external connections. Furthermore, they state that a Content Security Policy (CSP) in enforcement mode could have blocked key components of the attack.

Ultimately, the Optimizely incident reveals a critical gap in how we manage the software supply chain. As Report URI puts it: "trusted JavaScript doesn't always remain trustworthy." If your security strategy relies on the assumption that a vendor's CDN is an immutable source of truth, you aren't managing risk—you're just hoping you aren't the next snippet on the list.

Sources

More from Naomi Frost