Sentrix

Analysis · Payments

PCI DSS 6.4.3 and 11.6.1: the Optimizely test

The card skimmer served through Optimizely lasted less than a day. What PCI DSS requirements 6.4.3 and 11.6.1 would have caught, and what they leave out.

By Sentrix · Published 2026-11-17

On 16 November 2026, according to Report URI, code an attacker added to the Optimizely snippets tied to 178 organisations gained the ability to read payment cards typed into the sites' forms, wherever the snippet loaded on a card entry page. According to Report URI, the skimmer was added at 13:43 UTC, and every observed file was clean by 23:11 UTC. Less than 10 hours. Since 31 March 2025, PCI DSS has carried 2 requirements written for exactly this scenario. The full story of the incident is in our analysis of the Optimizely attack.

What the 2 requirements say

Requirement 6.4.3 covers every script loaded and executed in the browser on a payment page. It asks for 3 things: a method to confirm each script is authorised, a method to assure its integrity, and an inventory of all scripts with a written business or technical justification. Version 4.0.1 softened the wording, from a "necessary" script to a justified one, according to Report URI.

Requirement 11.6.1 asks for a change- and tamper-detection mechanism. It must alert personnel to unauthorised changes to the security-impacting HTTP headers and the script contents of payment pages, as received by the browser. The check runs at least weekly, or at the frequency set by a targeted risk analysis.

The PCI Security Standards Council confirms these requirements took effect on 31 March 2025. In March 2025 it published an information supplement on payment page security that opens with this finding: "Scripts running in consumers' browsers are now a significant target for attackers seeking to steal payment card data."

The Optimizely test, requirement by requirement

Inventory and authorisation. A merchant compliant with 6.4.3 had the Optimizely snippet in its inventory, justified and authorised. It was the right script, served by the right domain. The malicious code arrived inside it. Authorising a script says nothing about what it contains today.

Integrity. The simplest method, Subresource Integrity, pins a file's hash. It does not work for a snippet that changes with every experiment publish: StepSecurity notes such a file cannot be pinned to one hash. A merchant that keeps Optimizely on a payment page must assure integrity some other way. A Content Security Policy limits the domains other scripts can load from, but it does not check the content served by an allowed domain. A separate mechanism is needed: monitor the content served and its changes, or remove the script from the payment page.

Change detection. This is the requirement that could have raised the alarm. The snippet of every affected project changed, and a new script, track.js, appeared. But frequency matters: the card-skimming version was served for less than 10 hours, and the payload was rebuilt 26 times that day. A weekly check, which the text allows, could miss it entirely. More frequent monitoring, for example through CSP violation reports, could detect the change during the attack, depending on its configuration and the resources it watched.

The hosted payment frame. Report URI states that the code could not directly read the fields of a hosted payment frame, such as Stripe's or Adyen's. Cards typed into the site's own fields were within its reach. The payment architecture made more difference than any ticked box.

The SAQ A trap

In January 2025 the Council removed 6.4.3 and 11.6.1 from SAQ A, the questionnaire for merchants that fully outsource card entry to a provider. In exchange, it added an eligibility criterion: the merchant must confirm its site is not susceptible to attacks from scripts. The FAQ published in February 2025 offers 2 ways to do so: apply techniques such as those in 6.4.3 and 11.6.1, or obtain written confirmation from a compliant provider that its solution protects the page from such attacks.

The Council says it plainly: the change "does not remove or diminish the underlying requirements". An SAQ A merchant that loads Optimizely on the page holding the payment frame must therefore be able to show how it confirmed that criterion: through its own techniques, taking that script into account, or through its provider's written confirmation.

What the requirements do not cover

PCI DSS protects card data, on payment pages. The code served through Optimizely could run on every page that loaded the snippet: login, customer account, content. Depending on conditions, it could also collect cookies and screenshots, swap cryptocurrency addresses and try to exploit iPhones. None of that falls under 6.4.3 or 11.6.1. An organisation that only watches its payment pages saw, at best, part of the attack. The rest falls under personal information law, which we cover in the article on your duties under Law 25 and PIPEDA.

Our view at Sentrix

  1. Take experimentation scripts off payment pages. The surest way to pass 6.4.3 is to have few scripts to justify. A conversion experiment on the payment page is rarely worth the risk.
  2. Set the monitoring frequency on the real risk. The targeted risk analysis under requirement 12.3.1 should reflect what this incident shows: an attack can start and end between 2 weekly checks.
  3. Extend monitoring beyond payment. Login and account pages deserve the same inventory and monitoring, even though PCI DSS does not require it.
  4. Keep the evidence. The assessor will ask for the inventory, the justifications and the alerts. Our compliance automation ties that evidence to the PCI DSS requirements.

The next step

Open your 6.4.3 inventory and look for experimentation, analytics and marketing scripts on your payment pages. For each one, ask whether your 11.6.1 mechanism would have seen, on 16 November, a change that lasted less than a day. If you want to review these controls with us, contact us.

Sources

Frequently asked questions

Q01
What do requirements 6.4.3 and 11.6.1 ask for?
Requirement 6.4.3 asks, for every script loaded and executed on a payment page, for a method to confirm it is authorised, a method to assure its integrity and an inventory with a written justification. Requirement 11.6.1 asks for a mechanism that alerts on unauthorised changes to the security-impacting headers and the script contents of payment pages, as received by the browser.
Q02
Since when are they mandatory?
Since 31 March 2025, according to the PCI Security Standards Council. The Council removed these requirements from the SAQ A self-assessment questionnaire in January 2025, but added an eligibility criterion: the merchant must confirm its site is not susceptible to attacks from scripts. The Council states that this change does not remove or diminish the underlying requirements.
Q03
Would a weekly check have been enough?
Not for this incident. Requirement 11.6.1 allows a check at least weekly, or at the frequency set by a targeted risk analysis. According to Report URI, card skimming was added at 13:43 UTC on 16 November and every observed file was clean by 23:11 UTC. A weekly check could have missed the malicious version entirely.

Let's talk about your compliance program.

Last updated: