Another day, another round of “How could this happen?” questions asked by people who definitely did not read the last ten security advisories. Pour yourself something neat, maybe a scotch or a splash of rum, because the universe clearly decided subtlety is for defenders.
The top item making the rounds from this week’s security roundup is the story about attackers using Optimizely to A/B test malware on top websites, potentially impacting at least 178 sites over a 24-hour window. The basic premise: a vulnerability involving a third-party web experimentation platform was used to inject data-stealing malicious code into websites. Targets reportedly included major names such as BBC, CNN, Equifax, Ledger, and Zoom. (Because of course it did.)
Yes, It’s “Just Third-Party Content.” No, That’s Not a Security Plan
Here is the part that should make every security leader quietly stare into the middle distance: the incident was exacerbated by the fact that many popular websites do not enforce strict rules on third-party sources. In other words, the modern web is held together by trust, vibes, and a prayer written in JavaScript.
Attackers did not hack “the brand website” directly. They weaponized the plumbing – the vendor script ecosystem that teams bolt on because it’s convenient, cheap, and lets marketing run experiments without waiting for engineering. Vendors ship features. CISOs sign the paperwork. Then everyone acts surprised when the supply chain gets popped like a cork.
And let’s be honest: if you are still running unvetted third-party scripts in production, you are not “being agile.” You are just being optimistic in a way that only looks brave until it becomes expensive.
“A/B Test Malware” Should Be a Compliance Requirement, Not a Surprise
Attackers used the same mechanism intended for harmless experimentation to deliver malicious payloads. That’s the real horror movie setup. The technical details aside, the operational lesson is brutal: dynamic content injection via third-party platforms means your “perimeter” is more like a suggestion than a boundary.
So what should you do, besides filing tickets and scheduling a meeting titled “Third-Party Risk Strategy Q4”? Start treating web vendors like production dependencies with real security ownership: strict allowlists, content integrity controls, tighter CSP enforcement, monitoring for anomalous script behavior, and incident-ready playbooks that assume your vendor can become your threat.
Most orgs will probably do none of that. They will wait for the next newsletter line that says “x days ago, attackers exploited a third-party platform to inject skimmers and steal data.” Like it’s weather. Like it’s inevitable. Like your risk register is just decorative.
Read the Source (So You Can Pretend You Didn’t Miss It)
KrebsOnSecurity: Attackers used Optimizely to A/B test malware on top websites