PCI DSS 6.4.3 and 11.6.1: what the Council's open pages say about payment page scripts, and a ledger to keep
The PCI Security Standards Council says Requirements 6.4.3 and 11.6.1 of PCI DSS focus on the scripts on a payment page.
This guide reads them only from the Council's open web pages, its FAQs, glossary and blog. It did not open the standard, which the Council publishes in its Document Library, or any self-assessment questionnaire. It sets out what those pages say the requirements are for, when they took effect, how SAQ A changed and who decides, then gives you a script ledger and change-detection evidence sheet for your own payment page.
In this guide
Payment page script ledger and change-detection evidence sheet
A ledger in 17 rows and 6 columns, built from the PCI Security Standards Council's open pages. Rows L1 to L8 describe the payment page and each script on it: copy them into a spreadsheet as columns, one line per script, to make your script inventory.
How to use itHide the instructions
Rows C1 to C4 hold the change records: a baseline, each later comparison of scripts, frames, fields and headers, the alert review and a sign-off. Rows A1 to A5 are the SAQ A workflow: how the payment form reaches the customer, the written question to the acquirer or payment brand and, where the answer is SAQ A, the route, the provider's confirmation and the merchant's own confirmation. Rows L1, L2, L7, L8, C1 and C2 look at a page, and each carries its limits in its last column: your own site only, loaded as a customer would load it up to the payment step, with no card details entered, no payment submitted, and the payment provider's frame, fields and scripts recorded only as present, with the address each is served from. Only L1, C1 and C2 load pages; L2 and L7 work from L1's load, and L8 does too unless you use the security headers checker's site check, which sends one GET request from Clever Ops' server plus one for each redirect it follows, up to 5. Every other row works from records or puts a question to people in writing, and runs nothing against the payment page or any other system. The fourth column gives what the Council's open pages say, in their words or a close short form, naming the page each entry comes from (an FAQ by its number, a post by its date, the glossary or the Merchant Resources page); it says where a field is this guide's own and marks the entries that come from MDN or OWASP instead. The requirement text itself is in the standard. The ledger records evidence. Whether 6.4.3 or 11.6.1 is met, and whether SAQ A applies, is for the merchant with its acquirer or payment brand, and its assessor where one is engaged.
17 rows
Row and field
What to record
L1. The payment page and how it is built
- What to record
- The address of each page where a customer reaches the payment form; the address the payment form itself is served from; and which of the glossary's forms the page takes: a single document, a component in an inline frame within a non-payment page, or several components in several inline frames. If the customer is sent to the provider's site instead, record the redirect and where it goes.
- How to fill it in
- Walk through the checkout once with the browser's developer tools open. On the step where the payment form appears, note the address of each frame on the page from the Elements panel, without opening the frame itself.
- What the Council's open pages say it relates to
- PCI SSC glossary, Payment Page: a web-based user interface containing one or more form elements intended to capture account data from a consumer or submit captured account data, for purposes of processing and authorising payment transactions. FAQ 1588 (February 2025): the SAQ A script criterion only applies to merchants whose webpage includes a provider's embedded payment page or form.
- What the entry can show, and where it stops
- How the payment form reached one browser on one day. It does not settle whether a page is in scope for any requirement, or which questionnaire applies; row A2 takes that to the acquirer or payment brand.
- Limits of this row
- Limits for this row, which looks at a page: your own store only, on a staging copy where you have one and on the live site for the live record. Each check is limited to one product added to the cart, each page from the cart to the step where the payment form appears loaded once, at most one submission of checkout details, and at most one sign-in where the checkout needs a customer account. Repeat nothing and run no load testing. Any checkout details are invented, with an inbox your business controls, and any sign-in uses a test account holding invented details, not a real customer's account; warn whoever handles orders so the abandoned checkout is not taken for a customer's. Enter nothing in the payment fields, submit no payment and place no order. Record the payment provider's frame, fields and scripts as present, each with the address it is served from, and leave them alone: nothing inside them is opened, read, copied or run against. Load the pages in a fresh browser profile that was signed in to none of the business's systems when you opened it; the one test-account sign-in above, where the checkout needs it, is the only sign-in made in it. In the developer tools, keep the cookie and storage panels closed, read only the response headers this row names, if any, and never copy, paste or keep a Cookie or Set-Cookie value. Keep every password, key, token, cookie or session value, card number and link that works for whoever holds it out of the ledger. Should a page show a real customer's details, end the check at that page, write down only its address and the time, and hand the matter to whoever looks after privacy incidents in your business. In this row: the cart page and each checkout step up to the payment form, once each.
L2. Script address
- What to record
- For each script on the payment page: its full address, or "inline" and where it sits (the page template, a plugin's output, a tag); and whether it appears in the page source as well as in the network list. Ask your developer which script added any that appear only in the network list.
- How to fill it in
- On L1's load, filter the network panel to scripts and list every entry, then search the page source for script elements to find inline scripts. List the payment provider's own scripts with their addresses and mark them "provider", without opening them.
- What the Council's open pages say it relates to
- PCI SSC glossary, Payment Page Scripts: any programming language commands or instructions on a payment page that are processed and/or interpreted by a consumer's browser, including commands or instructions that interact with a page's document object model. FAQ 1581 (August 2024): the objective of 6.4.3 is to ensure that unauthorised code cannot be executed in the payment page as it is rendered in the consumer's browser.
- What the entry can show, and where it stops
- The scripts one load fetched, for one browser, on one build and day. If a script loads only for some visitors, devices or times, a single load may not show it, so note the browser, device and location you used.
- Limits of this row
- Limits for this row, which looks at a page: your own store only, on a staging copy where you have one and on the live site for the live record. Each check is limited to one product added to the cart, each page from the cart to the step where the payment form appears loaded once, at most one submission of checkout details, and at most one sign-in where the checkout needs a customer account. Repeat nothing and run no load testing. Any checkout details are invented, with an inbox your business controls, and any sign-in uses a test account holding invented details, not a real customer's account; warn whoever handles orders so the abandoned checkout is not taken for a customer's. Enter nothing in the payment fields, submit no payment and place no order. Record the payment provider's frame, fields and scripts as present, each with the address it is served from, and leave them alone: nothing inside them is opened, read, copied or run against. Load the pages in a fresh browser profile that was signed in to none of the business's systems when you opened it; the one test-account sign-in above, where the checkout needs it, is the only sign-in made in it. In the developer tools, keep the cookie and storage panels closed, read only the response headers this row names, if any, and never copy, paste or keep a Cookie or Set-Cookie value. Keep every password, key, token, cookie or session value, card number and link that works for whoever holds it out of the ledger. Should a page show a real customer's details, end the check at that page, write down only its address and the time, and hand the matter to whoever looks after privacy incidents in your business. In this row: no new loads; it works from the load recorded in L1.
L3. Source, and the account that can change it
- What to record
- Who serves each script: your own server, a named third party, the payment provider, or a tag manager, with the tag inside it that adds the script. For each, the account that can change what the script does or which scripts load.
- How to fill it in
- Work from L2's addresses and your own records: contracts, the plugin or app list, and the tags in the tag manager (ask the tag manager's owner whether it can export them). Ask your developer about any host nobody recognises.
- What the Council's open pages say it relates to
- FAQ 1592 (March 2025): a provider of third-party scripts is not a third-party service provider for Requirements 12.8 and 12.9, as part of an entity's assessment of its e-commerce environment, if the entity confirms that the provider's only service is providing scripts not related to payment processing and that its scripts cannot impact the security of cardholder data and/or sensitive authentication data. From OWASP's Third Party JavaScript Management Cheat Sheet, not a Council page: the content a tag manager returns can be dynamically changed by host site employees, such as the marketing part of the business.
- What the entry can show, and where it stops
- Who can change each script, as your records show it. It does not show whether a third party's server has itself been changed: OWASP's cheat sheet says there is usually no guarantee that the code hosted at the third party will remain the same as seen by the developers and testers.
- Limits of this row
- Read-only row: it works from L2's list and your own records and sends nothing to any website. Ask an account's owner for its list rather than signing in to it yourself.
L4. Purpose
- What to record
- One line on why each script is on the payment page, and which kind it is: payment, 3DS, fraud screening, analytics, advertising, chat, testing of page variants, or other.
- How to fill it in
- Ask whoever added it, and write down their answer and name. A script with no stated purpose goes to its owner in L5 as a question.
- What the Council's open pages say it relates to
- FAQ 1581: for merchants using a 3DS solution, validation to Requirement 6.4.3 for 3DS scripts is not required due to the inherent trust relationship between the 3DS service provider and the merchant, as established in the merchant's due diligence and onboarding processes and the business agreement between them; any script run outside the purpose of performing a 3DS functionality is subject to 6.4.3.
- What the entry can show, and where it stops
- Why the business says each script is there. Whether a reason justifies a script on the payment page is the merchant's call, and whether a script falls within FAQ 1581 is a question for the acquirer or assessor.
- Limits of this row
- Read-only row: it records answers from people and sends nothing to any website.
L5. Owner
- What to record
- A named person, not a shared inbox, who answers for each script and for any alert about it, with a named deputy.
- How to fill it in
- Agree it with the business, starting from the account holder in L3.
- What the Council's open pages say it relates to
- This field is this guide's own. The Council's podcast of 26 March 2025 says the requirements focus on payment page scripts being properly authorised, checked for integrity and monitored for tampering. Rows L6, L7 and C3 each need a person to ask.
- What the entry can show, and where it stops
- Who answers for a script. It says nothing about the script itself.
- Limits of this row
- Read-only row: it records names and sends nothing to any website.
L6. Authorisation
- What to record
- Who approved the script for the payment page, the date, and where the approval is kept: a ticket, an email or a change record. Where none is found, write "none found" and the date you asked.
- How to fill it in
- Search the change records for each script in L2. A script nobody approved and nobody recognises goes to C3 as an alert straight away, not to the next review.
- What the Council's open pages say it relates to
- Council podcast of 26 March 2025: the requirements focus on payment page scripts being properly authorised. FAQ 1581: the objective of 6.4.3, as set out in L2. What 6.4.3 asks is in its text in the standard.
- What the entry can show, and where it stops
- That an approval exists on paper. Whether it is what 6.4.3 asks for is judged against the requirement's text by the merchant and its assessor.
- Limits of this row
- Read-only row: it reads records and sends nothing to any website. Do not remove or change a script during the review: that is a change for its owner to make through the business's own change process.
L7. Integrity method
- What to record
- How the business knows each script is the one it approved: an integrity attribute on the script element, a copy hosted on your own server with its version and the date someone reviewed it, or none. For the payment provider's scripts, write "provider" and see A4.
- How to fill it in
- Read each external script element in the page source from L2 for an integrity attribute and a crossorigin attribute. For a self-hosted copy, ask who reviewed which version, and when.
- What the Council's open pages say it relates to
- Council podcast of 26 March 2025: checked for integrity. From MDN's Subresource Integrity page, not a Council page: a cryptographic hash that a fetched resource must match, a browser that refuses to load a resource whose hash matches none of the values in its integrity attribute, and cross-origin requests that must use CORS, so the crossorigin attribute must be included. From OWASP's cheat sheet: in-house script mirroring, Subresource Integrity and secure transmission of the third-party code are among typical defences.
- What the entry can show, and where it stops
- Whether the browser was told to check this script's content on this load. The check belongs to the element that carries it, so give any script that script adds its own line in L2.
- Limits of this row
- Limits for this row, which looks at a page: your own store only, on a staging copy where you have one and on the live site for the live record. Each check is limited to one product added to the cart, each page from the cart to the step where the payment form appears loaded once, at most one submission of checkout details, and at most one sign-in where the checkout needs a customer account. Repeat nothing and run no load testing. Any checkout details are invented, with an inbox your business controls, and any sign-in uses a test account holding invented details, not a real customer's account; warn whoever handles orders so the abandoned checkout is not taken for a customer's. Enter nothing in the payment fields, submit no payment and place no order. Record the payment provider's frame, fields and scripts as present, each with the address it is served from, and leave them alone: nothing inside them is opened, read, copied or run against. Load the pages in a fresh browser profile that was signed in to none of the business's systems when you opened it; the one test-account sign-in above, where the checkout needs it, is the only sign-in made in it. In the developer tools, keep the cookie and storage panels closed, read only the response headers this row names, if any, and never copy, paste or keep a Cookie or Set-Cookie value. Keep every password, key, token, cookie or session value, card number and link that works for whoever holds it out of the ledger. Should a page show a real customer's details, end the check at that page, write down only its address and the time, and hand the matter to whoever looks after privacy incidents in your business. In this row: no new loads; it reads the page source captured in L2.
L8. Content Security Policy coverage
- What to record
- Whether the payment page's response carried Content-Security-Policy, Content-Security-Policy-Report-Only, both or neither; for each script host in L2, whether the policy lists it in script-src-elem where that directive is present, otherwise in script-src, or in default-src where neither is present; and whether the policy names a reporting endpoint, with the person who reads what arrives there.
- How to fill it in
- Work from L1's load, taken in the fresh browser profile the limits describe. In the developer tools, read only the Content-Security-Policy and Content-Security-Policy-Report-Only headers of the payment step's response, and record those two by name and value. Never copy, paste or keep a Cookie or Set-Cookie value. To have the security headers checker's paste tab analyse them in your browser, paste in only those two headers. The checker's site check reads one public address without signing in, so it may not reach a checkout that needs a cart or a sign-in.
- What the Council's open pages say it relates to
- Council post of 10 March 2025: the requirements also address the need for careful management of both scripts and security-impacting HTTP headers to prevent unauthorised changes to webpages, whether in the server environment or as rendered in the consumer's browser. From MDN, not a Council page: the script-src directive sets the allowed sources for JavaScript and is the fallback for script-src-elem, default-src is the fallback for the other fetch directives, and a policy sent as Content-Security-Policy-Report-Only is not enforced but sends violations to the reporting endpoint the policy names.
- What the entry can show, and where it stops
- Which hosts the policy allowed on this response, and whether it was enforced. It does not show what those hosts serve, and a report-only policy is not enforced.
- Limits of this row
- Limits for this row, which looks at a page: your own store only, on a staging copy where you have one and on the live site for the live record. Each check is limited to one product added to the cart, each page from the cart to the step where the payment form appears loaded once, at most one submission of checkout details, and at most one sign-in where the checkout needs a customer account. Repeat nothing and run no load testing. Any checkout details are invented, with an inbox your business controls, and any sign-in uses a test account holding invented details, not a real customer's account; warn whoever handles orders so the abandoned checkout is not taken for a customer's. Enter nothing in the payment fields, submit no payment and place no order. Record the payment provider's frame, fields and scripts as present, each with the address it is served from, and leave them alone: nothing inside them is opened, read, copied or run against. Load the pages in a fresh browser profile that was signed in to none of the business's systems when you opened it; the one test-account sign-in above, where the checkout needs it, is the only sign-in made in it. In the developer tools, keep the cookie and storage panels closed, read only the response headers this row names, if any, and never copy, paste or keep a Cookie or Set-Cookie value. Keep every password, key, token, cookie or session value, card number and link that works for whoever holds it out of the ledger. Should a page show a real customer's details, end the check at that page, write down only its address and the time, and hand the matter to whoever looks after privacy incidents in your business. In this row: no new loads; it reads the headers of L1's load. If you use the checker's site check on a public page instead, it sends one GET request from Clever Ops' server, plus one for each redirect it follows, up to 5, so give it only an address on your own site.
C1. Baseline record
- What to record
- The date, time, build or release, browser and device; the full L2 list with each script's L7 and L8 entries; the payment page's response headers, by name and value where present, limited to these and no others: Content-Security-Policy, Content-Security-Policy-Report-Only, Strict-Transport-Security, X-Frame-Options, X-Content-Type-Options, Referrer-Policy and Permissions-Policy; every frame on the payment step by its address; and every form and input field outside the provider's frame, by the name or id each carries.
- How to fill it in
- Take it once L1 to L8 are complete, from one new load in the fresh browser profile the limits describe, reading only the headers this row lists and copying, pasting or keeping no Cookie or Set-Cookie value, and have the reviewer named in C4 sign it. Keep it where it cannot be edited without a trace, such as a dated file with version history.
- What the Council's open pages say it relates to
- Council podcast of 26 March 2025: monitored for tampering. Council post of 10 March 2025: careful management of both scripts and security-impacting HTTP headers to prevent unauthorised changes to webpages, whether in the server environment or as rendered in the consumer's browser. Neither page says which of 6.4.3 and 11.6.1 covers which aim; this row is this guide's own design, and the requirement text is in the standard.
- What the entry can show, and where it stops
- What the page looked like to one browser on one day. It is the reference C2 compares against, not evidence that nothing was wrong that day.
- Limits of this row
- Limits for this row, which looks at a page: your own store only, on a staging copy where you have one and on the live site for the live record. Each check is limited to one product added to the cart, each page from the cart to the step where the payment form appears loaded once, at most one submission of checkout details, and at most one sign-in where the checkout needs a customer account. Repeat nothing and run no load testing. Any checkout details are invented, with an inbox your business controls, and any sign-in uses a test account holding invented details, not a real customer's account; warn whoever handles orders so the abandoned checkout is not taken for a customer's. Enter nothing in the payment fields, submit no payment and place no order. Record the payment provider's frame, fields and scripts as present, each with the address it is served from, and leave them alone: nothing inside them is opened, read, copied or run against. Load the pages in a fresh browser profile that was signed in to none of the business's systems when you opened it; the one test-account sign-in above, where the checkout needs it, is the only sign-in made in it. In the developer tools, keep the cookie and storage panels closed, read only the response headers this row names, if any, and never copy, paste or keep a Cookie or Set-Cookie value. Keep every password, key, token, cookie or session value, card number and link that works for whoever holds it out of the ledger. Should a page show a real customer's details, end the check at that page, write down only its address and the time, and hand the matter to whoever looks after privacy incidents in your business. In this row: the same pages as L1, once each.
C2. Observed DOM and header changes against the baseline
- What to record
- For each later check: the date, time, build and every difference from C1. That means a script added, removed or served from a new address; an integrity attribute added, removed or changed; one of the headers C1 lists added, removed or changed; a frame added or served from a new address; and a form or input field added outside the provider's frame. Where nothing differs, write "no difference seen on this load".
- How to fill it in
- Repeat C1's load the same way and compare the lists line by line. Where the business runs a change-detection mechanism of its own, its host's or a provider's, put what it reported for the same period beside your comparison. This ledger sets no schedule for this row: read 11.6.1 and 12.3.1 in the standard, and ask your acquirer or assessor.
- What the Council's open pages say it relates to
- Council post of 30 January 2025: Requirement 12.3.1 is for a targeted risk analysis to support Requirement 11.6.1. The text of both is in the standard. This row is this guide's own design; none of the open Council pages read for this guide says what 11.6.1 asks.
- What the entry can show, and where it stops
- A difference between two loads in one browser. It cannot show a change that came and went between checks, or one served only to some visitors, which is why C3 also reviews whatever mechanism the business runs.
- Limits of this row
- Limits for this row, which looks at a page: your own store only, on a staging copy where you have one and on the live site for the live record. Each check is limited to one product added to the cart, each page from the cart to the step where the payment form appears loaded once, at most one submission of checkout details, and at most one sign-in where the checkout needs a customer account. Repeat nothing and run no load testing. Any checkout details are invented, with an inbox your business controls, and any sign-in uses a test account holding invented details, not a real customer's account; warn whoever handles orders so the abandoned checkout is not taken for a customer's. Enter nothing in the payment fields, submit no payment and place no order. Record the payment provider's frame, fields and scripts as present, each with the address it is served from, and leave them alone: nothing inside them is opened, read, copied or run against. Load the pages in a fresh browser profile that was signed in to none of the business's systems when you opened it; the one test-account sign-in above, where the checkout needs it, is the only sign-in made in it. In the developer tools, keep the cookie and storage panels closed, read only the response headers this row names, if any, and never copy, paste or keep a Cookie or Set-Cookie value. Keep every password, key, token, cookie or session value, card number and link that works for whoever holds it out of the ledger. Should a page show a real customer's details, end the check at that page, write down only its address and the time, and hand the matter to whoever looks after privacy incidents in your business. In this row: the same pages as C1, once each per check.
C3. Alert review
- What to record
- For each difference from C2 and each alert from the business's mechanism, including Content Security Policy violation reports where the policy names an endpoint: when it arrived, who received it, who reviewed it and when, what it turned out to be (an approved release, a vendor's own update, or unexplained), and what was done: L2 to L8 updated, the change reversed by its owner, or an incident opened.
- How to fill it in
- Work from the mechanism's inbox and the change records. Hand anything unexplained on the payment page to whoever handles incidents straight away.
- What the Council's open pages say it relates to
- Council podcast of 26 March 2025: monitored for tampering. From MDN, not a Council page: the recommended method for reporting CSP violations is the Reporting API with the report-to directive; report-uri is deprecated, but MDN says to declare both until report-to is supported in all browsers. This row is this guide's own design; none of the open Council pages read for this guide says what 11.6.1 asks.
- What the entry can show, and where it stops
- That a named person looked at each alert, and what they decided. Whether the mechanism and its review meet 6.4.3, 11.6.1 or neither is judged against the requirements' text by the merchant and its assessor.
- Limits of this row
- Records row: it reads alerts and records decisions, and runs nothing against any website. A violation report can carry page addresses; keep tokens, session values and customer details out of the copy you file, and pass any report that holds a real customer's details to whoever looks after privacy incidents in your business.
C4. Review and sign-off
- What to record
- Who reviewed the ledger, the date, the build, what the review covered (pages, loads and rows) and what it did not, and each open question with its owner and a date.
- How to fill it in
- The person the business names for this signs after C1, and after each round of C2 and C3.
- What the Council's open pages say it relates to
- This row is this guide's own. The Council's post of 30 January 2025 says compliance validation requirements are set by brands, acquirers, payment facilitators and others, often referred to as compliance enforcing entities.
- What the entry can show, and where it stops
- That the business reviewed its own record on that date. It is not a statement that 6.4.3 or 11.6.1 is met, that the page is free of skimming code, or that any questionnaire applies.
- Limits of this row
- Read-only row: it signs a record and sends nothing to any website.
A1. How the payment form reaches the customer
- What to record
- From L1: whether the customer pays in the provider's payment page or form embedded in your page (one or more iframes), is redirected to the provider (an HTTP 30x redirect, a meta redirect tag or a JavaScript redirect), pays through a link to the provider's website such as one sent by email, or types card details into a form your own site serves.
- How to fill it in
- Copy the answer and its evidence from L1.
- What the Council's open pages say it relates to
- FAQ 1588 (February 2025): the SAQ A script criterion only applies to e-commerce merchants with a webpage that includes a provider's embedded payment page or form, and does not apply to merchants that redirect customers to the provider or fully outsource payment functions.
- What the entry can show, and where it stops
- How payment was delivered on L1's load. Which questionnaire applies, if any, is not decided here.
- Limits of this row
- Read-only row: it works from L1's record and sends nothing to any website.
A2. The question to your acquirer or payment brand
- What to record
- Your written question asking whether you must submit a self-assessment questionnaire and, if so, which one, with A1's record attached; their written answer and its date.
- How to fill it in
- Send it to the entity you would submit the questionnaire to. Where you work with an assessor, put the same question to them.
- What the Council's open pages say it relates to
- FAQ 1588: merchants should consult their compliance-accepting entity, the entity to which the SAQ will be submitted (typically, an acquirer (merchant bank) or the payment brands), to determine whether they are required to submit an SAQ and, if so, which SAQ is appropriate. Merchant Resources page: for questions about compliance validation and reporting requirements, merchants should contact their acquirer (merchant bank) or the payment brand they do business with, as applicable.
- What the entry can show, and where it stops
- Their answer, in writing, with its date. The ledger holds it and does not second-guess it.
- Limits of this row
- Records row: it puts a written question to people and runs nothing against any website or system.
A3. The route to the script criterion
- What to record
- Only where A2's answer is SAQ A and A1 shows an embedded payment page or form: which of FAQ 1588's two routes the merchant relies on, and the evidence held for it.
- How to fill it in
- Route one: techniques such as those detailed in 6.4.3 and 11.6.1, deployed by the business or a third party, with rows L1 to C4 kept up to date as the record. Route two: the provider's written confirmation described in A4.
- What the Council's open pages say it relates to
- FAQ 1588: the merchant can confirm the criterion either by using techniques such as, but not limited to, those detailed in Requirements 6.4.3 and 11.6.1 to protect the merchant's webpage from scripts targeting account data, which the merchant or a third party may deploy; or by obtaining confirmation from its PCI DSS compliant provider of the embedded payment page or form that, when implemented according to the provider's instructions, the provider's solution includes techniques that protect the merchant's payment page from script attacks.
- What the entry can show, and where it stops
- Which route the merchant relies on and what it holds. Whether that is enough is for the acquirer or payment brand.
- Limits of this row
- Read-only row: it records a choice and its documents and sends nothing to any website.
A4. The provider's confirmation and instructions
- What to record
- For route two: the provider's confirmation, its date, the implementation instructions it refers to with their version or date, and your developer's dated note of how your payment page follows those instructions. For either route: the provider's guidance on implementing its solution securely.
- How to fill it in
- Ask the provider in writing. Do not test the provider's frame, fields, scripts or platform to check what it says; Clever Ops does not test a payment provider's own product or platform either.
- What the Council's open pages say it relates to
- FAQ 1588: the confirmation is that, when implemented according to the provider's instructions, its solution includes techniques that protect the merchant's payment page from script attacks; merchants are encouraged to work with their provider to obtain guidance about how to implement its solution securely.
- What the entry can show, and where it stops
- What the provider stated, and how your developer says the page follows it. The ledger does not verify the provider's statement.
- Limits of this row
- Records row: it puts a written question to the provider, collects documents, and runs nothing against the provider's frame, fields, scripts or platform or any other system.
A5. The merchant's own confirmation
- What to record
- Who in the business made the confirmation the criterion describes, the date, the route from A3, the rows it relied on, and what will make the business look at it again, such as a C2 difference, a new provider or a new script.
- How to fill it in
- The merchant makes and records this confirmation itself. A tester, a developer or this ledger does not make it on the merchant's behalf.
- What the Council's open pages say it relates to
- FAQ 1588, giving the criterion as PCI DSS v4.0.1 SAQ A r1 words it: "The merchant has confirmed that their site is not susceptible to attacks from scripts that could affect the merchant's e-commerce system(s)."
- What the entry can show, and where it stops
- The merchant's own statement, with its evidence listed. What the acquirer or payment brand accepts is for them.
- Limits of this row
- Read-only row: it records a decision and sends nothing to any website.
PCI DSS 6.4.3 and 11.6.1: what the Council's open pages say they are for
The PCI Security Standards Council (PCI SSC) publishes PCI DSS, which its Merchant Resources page calls a set of baseline technical and operational requirements designed to protect payment account data.
The same page says PCI DSS is intended for all entities that store, process or transmit cardholder data and/or sensitive authentication data, or could impact the security of that data, and that whether a small merchant is required to validate compliance is determined by the individual payment brands.
Whether and how it applies to your business, and how you validate compliance, are questions to put to your acquirer or payment brand, as the section on who decides sets out.
This guide covers Requirements 6.4.3 and 11.6.1, which a Council post of 30 January 2025 calls the new e-commerce security requirements in PCI DSS v4.0.1, and the SAQ A eligibility criterion about scripts that the same post announced.
The standard as a whole, self-assessment questionnaires in general, scans by an Approved Scanning Vendor, penetration testing, scope, segmentation and tester independence are the ground of the PCI DSS compliance guide in this library.
The wording of both requirements is in the standard itself, which the Council publishes in its Document Library. The Council's privacy policy speaks of users downloading materials from its website and executing corresponding licences. This guide opened neither the library nor any licence, and it found no open Council page that sets out either requirement in full.
What follows is what the Council says about them on its open web pages (its FAQs, glossary, Merchant Resources and assessor pages, and PCI Perspectives blog), with MDN and OWASP named wherever the guide draws on them instead.
Before you rely on a requirement, obtain the standard from the Council and read the requirement's text, with any applicability notes, there.
FAQ 1581, dated August 2024, states the purpose of 6.4.3 directly: "The objective of PCI DSS Requirement 6.4.3 is to ensure that unauthorized code cannot be executed in the payment page as it is rendered in the consumer's browser." The rest of that FAQ deals with 3DS scripts, which the section on payment page scripts below takes up.
For the two requirements together, the open pages this guide read include the transcript of a Council podcast posted on 26 March 2025. In it, the Council's Director of Data Security Standards says 6.4.3 and 11.6.1 were added to PCI DSS v4.0, and appear in v4.0.1, to reduce the risk of e-skimming attacks, which the episode describes as data breaches during e-commerce transactions.
The Director also says scripts running in a consumer's browser are "now a significant target for attackers seeking to steal payment card data", and that the requirements focus on making sure payment page scripts are properly authorised, checked for integrity and monitored for tampering, and on preventing unauthorised changes to web pages.
The Council's post of 10 March 2025, which announced its information supplement "Payment Page Security and Preventing E-Skimming", brings in HTTP headers as well: it says the requirements also address the need for careful management of both scripts and security-impacting HTTP headers to prevent unauthorised changes to webpages, whether in the server environment or as rendered in the consumer's browser.
The open pages read for this guide say less about 11.6.1 on its own. The Council's post of 30 January 2025 refers to Requirement 12.3.1 for a targeted risk analysis to support Requirement 11.6.1, so read the two side by side in the standard.
Those pages also do not say which of the aims above falls under which requirement. The ledger's change records in rows C1 to C4 are this guide's own design, tied to those descriptions of both requirements together, not to the text of either.
Which edition of PCI DSS, and when 6.4.3 and 11.6.1 took effect
Every version number and date in this section comes from a dated Council post or FAQ, named with its date.
The Council's post of 11 June 2024 introduced PCI DSS v4.0.1 as a limited revision, published to address stakeholder feedback and questions received since v4.0 came out in March 2022. By the post's account, it fixes formatting and typographical errors, clarifies the focus and intent of some requirements and guidance, and adds or deletes no requirements.
One of the changes it lists under Requirement 6 added Applicability Notes to clarify how the requirement for managing payment page scripts applies.
It gave 31 December 2024 as the retirement date for v4.0, after which v4.0.1 would be the only active version of the standard the Council supports.
In the March 2025 podcast, the Director says both requirements are among PCI DSS's future-dated requirements, that these had been in the standard since March 2022 as "best practices" until 31 March 2025, and that after that date they are required and must be fully considered during a PCI DSS assessment.
The post of 30 January 2025 says PCI DSS v4.0.1 Requirements 6.4.3, 11.6.1 and 12.3.1 become effective as of 31 March 2025. A Council post of 28 February 2025, which announced FAQ 1588, says new requirements under PCI DSS v4.0.1 take effect on 1 April 2025; this guide follows the 30 January post's wording.
FAQ 1585, dated October 2024, covers the reporting side: a future-dated requirement an entity has not yet implemented may be marked Not Applicable in a Report on Compliance or self-assessment questionnaire completed before its effective date, and from that date all requirements applicable to the entity's assessment, including the newly effective ones, must be fully considered.
A Council post dated 3 June 2026 invited eligible stakeholders to review and give feedback on the published PCI DSS v4.0.1 during a six-week request for comments, from 3 June to 20 July. The post says the Council is beginning the next iteration of the standard by soliciting industry feedback, and that participants are required to accept a non-disclosure agreement to download the document.
This guide names v4.0.1 because the dated pages above do. Before you rely on a requirement number here, check the Council's site for a later version, and use the edition your acquirer or assessor names.
What the Council counts as a payment page and a payment page script
The Council's glossary entries for a payment page and for payment page scripts set the edges of the ledger.
The glossary defines a payment page as a web-based user interface containing one or more form elements intended to capture account data from a consumer or submit captured account data, for purposes of processing and authorising payment transactions. It says the page can be rendered as a single document or instance, as a document or component displayed in an inline frame within a non-payment page, or as several documents or components, each with one or more form elements, in several inline frames within a non-payment page.
It defines payment page scripts as any programming language commands or instructions on a payment page that are processed and/or interpreted by a consumer's browser, including commands or instructions that interact with a page's document object model. Its examples of programming languages are JavaScript and VB script, and it says neither markup languages such as HTML nor style rules such as CSS are programming languages.
The definition does not say where a script is served from, so row L2 of the ledger lists scripts written into the page beside scripts fetched from elsewhere.
FAQs 1581 and 1592 deal with scripts that come from other businesses. FAQ 1581 says that for merchants using a 3DS solution, validation to Requirement 6.4.3 for 3DS scripts is not required, due to the inherent trust relationship between the 3DS service provider and the merchant, as established in the merchant's due diligence and onboarding processes and the business agreement between them.
It adds that any script run outside of the purpose of performing a 3DS functionality is subject to Requirement 6.4.3. FAQ 1592, dated March 2025, answers a different question: whether a script provider counts as a third-party service provider for Requirements 12.8 and 12.9.
It says it does not, as part of an entity's assessment of its e-commerce environment, if the entity confirms both that the provider's only service is providing scripts not related to payment processing and that those scripts cannot impact the security of cardholder data and/or sensitive authentication data.
FAQ 1592 is about Requirements 12.8 and 12.9, not 6.4.3. Row L4 records each script's purpose so that questions about a particular script reach your acquirer or assessor with the facts attached.
Where the provider's frame sits
MDN describes the iframe element as a nested browsing context that embeds another document into the current one, and says script access to a frame's content is subject to the same-origin policy. OWASP's Third Party JavaScript Management Cheat Sheet describes putting vendor JavaScript into an iframe from a different domain, so that the vendor code has no direct access to the host page's DOM and cookies.
Those statements are about access across the frame's edge. For the page around the frame, MDN's guide to Content Security Policy says code an attacker manages to run in the context of a website can do anything the website's own code could do, its examples including accessing or modifying the content of the site's loaded pages.
The page around the frame is the merchant's, and FAQ 1588 speaks of protecting the merchant's webpage from scripts targeting account data. That surrounding page is what the ledger records.
SAQ A script requirements: removed from the questionnaire, and an eligibility criterion added
SAQ A is one of the Council's self-assessment questionnaires.
The post of 30 January 2025 says it includes only those PCI DSS requirements applicable to merchants with account data functions completely outsourced to PCI DSS validated and compliant third parties, where the merchant retains only paper reports or receipts with account data.
It says SAQ A merchants may be e-commerce or mail and telephone order merchants, and do not store, process or transmit any account data in electronic form on their systems or premises.
The same post announced changes to SAQ A in response to stakeholder feedback about the complexity of implementing 6.4.3 and 11.6.1. They removed Requirements 6.4.3 and 11.6.1, and Requirement 12.3.1 for the targeted risk analysis that supports 11.6.1, and added an eligibility criterion for merchants to "confirm their site is not susceptible to attacks from scripts that could affect the merchant's e-commerce system(s)." The post says the October 2024 version of SAQ A would be retired on 31 March 2025, the date from which the January 2025 version was to take effect.
It also says the changes affect how merchants approach compliance reporting for these requirements but do not remove or diminish the underlying requirements within PCI DSS, and in the March 2025 podcast the Director says the requirements were only removed from SAQ A and are still in the standard.
FAQ 1588, dated February 2025, gives the criterion in the words of PCI DSS v4.0.1 SAQ A r1: "The merchant has confirmed that their site is not susceptible to attacks from scripts that could affect the merchant's e-commerce system(s)." A note under it sends readers to SAQ A itself for the complete list of eligibility criteria.
The FAQ says the criterion only applies to e-commerce merchants whose webpage includes a third-party service provider's or payment processor's embedded payment page or form, for example in one or more iframes.
It does not apply to merchants whose webpage redirects customers to the provider, with examples including an HTTP 30x redirect, a meta redirect tag or a JavaScript redirect, or to merchants that fully outsource payment functions, for example by emailing customers a link to the provider's website to pay.
The FAQ also says it is only intended to clarify this criterion, and should not be interpreted to impact or contradict any other eligibility criteria in SAQ A or in any other SAQ.
Under FAQ 1588, the merchant can confirm the criterion by either of two routes:
- Using techniques such as, but not limited to, those detailed in Requirements 6.4.3 and 11.6.1 to protect the merchant's webpage from scripts targeting account data. The FAQ says the merchant or a third party may deploy them.
- Obtaining confirmation from the merchant's PCI DSS compliant third-party service provider or payment processor that provides the embedded payment page or form that, when implemented according to that provider's instructions, its solution includes techniques that protect the merchant's payment page from script attacks.
Who FAQ 1588 sends merchants to
FAQ 1588 encourages merchants to work with their provider to obtain guidance on implementing the provider's solution securely. It says merchants should continue to consult their compliance-accepting entity, the entity to which the SAQ will be submitted (typically, an acquirer (merchant bank) or the payment brands), to determine whether they are required to submit an SAQ and, if so, which one is appropriate for their environment.
It also puts the point about script providers in SAQ A terms: a provider of third-party scripts is not a third-party service provider for purposes of SAQ A if its only service is providing scripts not related to payment processing and those scripts cannot impact the security of cardholder data and/or sensitive authentication data.
Rows A1 to A5 of the ledger turn the FAQ into steps: how the payment form reaches the customer, the written question to the acquirer or payment brand, the route the merchant relies on, the provider's confirmation where that is the route, and the merchant's own confirmation. At each step the decision stays with the merchant and the entity it reports to.
What a review of your own payment page can show, and what it cannot
The rows that look at the page (L1, L2, L7, L8, C1 and C2) record what one browser received on the loads they name, on a stated build and day.
Within that, a review can record which scripts the browser fetched and from where, which external scripts carried an integrity check, which security headers came with the page, and what changed since the last recorded load. Subresource Integrity and Content Security Policy carry much of that record. MDN says Subresource Integrity lets a browser verify that a resource it fetches is delivered without unexpected manipulation, by providing a cryptographic hash the resource must match, and that the browser refuses to load a resource whose hash matches none of the values in its integrity attribute.
MDN gives the primary use of a Content Security Policy as controlling which resources, in particular JavaScript, a document is allowed to load, with the script-src directive setting the allowed sources for JavaScript, and says a policy sent as Content-Security-Policy-Report-Only is not enforced but sends violations to the reporting endpoint the policy names.
Those records are evidence about the page. Several things sit outside them:
- What other visitors received. If a script loads only for some visitors, devices or times, one load may not show it. OWASP's cheat sheet also says the content a tag manager returns can be dynamically changed by host site employees, such as the marketing part of the business.
- What happens on another company's server. The same OWASP cheat sheet says "there is usually no guarantee that the code hosted at the third-party will remain the same as seen from the developers and testers".
- What runs inside the payment provider's frame or fields. The ledger records that they are present and where they are served from, and goes no further.
- Whether a script's code is harmful. The ledger records what loads, who approved it and whether it changed, not what each script does.
- Whether the site is "not susceptible" to script attacks. That is the confirmation FAQ 1588 has the merchant make, by one of its two routes.
- Whether 6.4.3 or 11.6.1 is met, or whether SAQ A is the right questionnaire. Those are judged against the standard's text by the merchant, its acquirer or payment brand, and its assessor where one is engaged.
How to fill in the payment page script ledger
Whoever runs the online shop day to day, with the developer or agency behind the site, can fill in the ledger on this page.
Its rows fall into three groups. Rows L1 to L8 describe the payment page and each script on it: copy them into a spreadsheet as columns, with one line per script, and they become a payment page script inventory.
Rows C1 to C4 are the change-detection evidence sheet, with a dated baseline, each later comparison, the review of every alert and a sign-off. Rows A1 to A5 are the SAQ A workflow: how the payment form reaches the customer, the written question to the acquirer or payment brand and, where the answer is SAQ A, the route, the provider's confirmation and the merchant's own confirmation.
- Start with L1 and A1, which record how the payment form reaches the customer. A store that redirects customers to its provider, or fully outsources payment, for example through an emailed link to the provider's website, is one FAQ 1588 says the script criterion does not apply to, and its acquirer or payment brand is still the one to ask which questionnaire, if any, applies.
- Rows L1, L2, L7, L8, C1 and C2 look at the page, and each carries its limits in its last column: your own site only, loaded as a customer would load it up to the step where the payment form appears, with no card details entered and no payment submitted. Only L1, C1 and C2 load pages, and L8 sends a request only if you use the security headers checker's site check. The other rows work from records, from L1's load or from written questions to people, and run nothing against the payment page or any other system.
- Treat the payment provider's frame, fields and scripts as the provider's. Record that each is present and the address it is served from, and leave everything inside them alone.
- Record no password, key, token, cookie or session value, card number, or any link that opens something for whoever has it, anywhere in the ledger.
- If a comparison in C2 turns up a script, frame or form field nobody can account for, do not wait for the next review. Hand it to whoever handles incidents in your business at once, and tell the payment provider through the contact route it gives merchants. For the first hour after that, see the hacked website guide in this library.
- This page sets no schedule for C2. Read 11.6.1 and 12.3.1 in the standard, and ask your acquirer or assessor what they expect to see.
- For row L8, the security headers checker can read a page's headers. Its site check sends one GET request from Clever Ops' server to a public address, plus one for each redirect it follows, up to 5, and never signs in. For a checkout that needs a cart or a sign-in, read the Content-Security-Policy and Content-Security-Policy-Report-Only headers in your browser's developer tools, in the fresh browser profile the row's limits describe, and paste only those two into its paste tab instead, never a Cookie or Set-Cookie value: the paste tab analyses them in your browser and sends nothing. Check only a site you own or have permission to check.
Who decides whether 6.4.3, 11.6.1 or SAQ A applies
The Council's open pages say it neither sets validation responsibilities nor determines whether a particular implementation is compliant.
Its post of 30 January 2025 says PCI SSC does not define compliance requirements for any organisation or set compliance validation responsibilities, that compliance validation requirements are set by brands, acquirers, payment facilitators and others, which it says are often referred to as compliance enforcing entities, and that organisations must consult their compliance enforcing entity with questions about validation requirements or the validation tools they may be eligible to use.
Its post of 10 March 2025 says the Council does not enforce compliance or determine whether specific implementations are compliant. Its glossary describes an acquirer, also called a merchant bank, as an entity, typically a financial institution, that processes payment card transactions for merchants and is defined by a payment brand as an acquirer, and says acquirers are subject to payment brand rules and procedures regarding merchant compliance.
For a merchant, the open pages point to the following parties. The merchant makes the confirmation the SAQ A criterion describes, and its own records, this ledger among them, can hold the evidence it relies on. The compliance-accepting entity, typically an acquirer or the payment brands, is the one FAQ 1588 sends merchants to, to determine whether they submit a questionnaire and which one, and the Merchant Resources page says that for questions about compliance validation and reporting requirements, merchants should contact the acquirer or payment brand they do business with, as applicable.
Where an assessor is engaged, the Council describes Qualified Security Assessor companies as independent security organisations qualified by the Council to validate an entity's adherence to PCI DSS.
Ask your acquirer, and your assessor if you have one, in writing which questionnaire or report applies to you, and keep the answer in row A2.
Neither a test nor a filled-in ledger makes that decision. Each records what a page did on a given day, and neither makes a merchant compliant with PCI DSS, eligible for SAQ A, or free of skimming code.
Where a Clever Ops test fits beside 6.4.3 and 11.6.1
The services Clever Ops offers test what a customer, or an attacker, can get to through a browser or an API.
On a store, the WordPress security audit and the web application and API penetration test can reach the checkout, and the website security check reads what the public pages send. For a WooCommerce store, the WordPress security audit covers the cart and checkout requests, on a staging copy or with the gateway in test mode, and lists each script the checkout page loads, which shows whose code runs where customers enter their details.
For the build the audit covered, that list is a starting point for row L2. For a custom-built store with customer sign-in, the web application and API penetration test takes in payment and billing flows on the application's side of the payment, run on a staging copy tied to a non-live payment account if one is available.
The website security check, run from outside without signing in, reads the headers and cookies on public responses.
A report sets out the scope as tested, then each finding with its evidence (secrets redacted) and a fix, and the one consolidated retest that comes with every test goes back over the reported findings within the agreed scope. What the merchant gets is evidence to keep with its own records and to show the people it reports to.
A report is not a PCI DSS assessment, it does not confirm the SAQ A criterion on anyone's behalf, and it is not an Approved Scanning Vendor scan.
Other parts of the work belong elsewhere. Clever Ops is neither a PCI Approved Scanning Vendor nor a Qualified Security Assessor, so a store that needs an external scan or an assessment in one of those roles is referred to a company the Council lists in that role.
Clever Ops does not test a payment provider's own product or platform, so the provider's hosted pages and platform stay untested, and the ledger leaves them alone too. Round-the-clock monitoring is not a Clever Ops service either, so any change-detection mechanism whose reports reach rows C2 and C3 is one the business or its providers run.
When Clever Ops is the store's own builder or host, its work there goes by the name security review, not independent test, and its report states the business that built the store and the business that hosts it.
Frequently asked questions
Its full text is in the standard, which the Council publishes in its Document Library and this guide did not open. On its open pages, the Council's FAQ 1581 says the objective of 6.4.3 is to ensure that unauthorised code cannot be executed in the payment page as it is rendered in the consumer's browser, and its March 2025 podcast says 6.4.3 and 11.6.1 focus on payment page scripts being properly authorised, checked for integrity and monitored for tampering. Ledger rows L2 to L8, with C1 to C3 for monitoring, record facts that relate to those descriptions, and your acquirer or assessor weighs them against the requirement's text.
The Council's post of 30 January 2025 names 11.6.1 with 6.4.3 as the new e-commerce security requirements in PCI DSS v4.0.1. The open Council pages read for this guide do not set out its wording. The same post says 6.4.3, 11.6.1 and 12.3.1 become effective as of 31 March 2025 (a Council post of 28 February 2025 says new requirements under v4.0.1 take effect on 1 April 2025), and refers to Requirement 12.3.1 for a targeted risk analysis to support 11.6.1. Read both in the standard. The ledger's change-detection rows, C1 to C4, are this guide's own design for keeping change records, not the requirement's text.
The Council's post of 30 January 2025 announced the removal of both requirements from SAQ A and the addition of an eligibility criterion about scripts, and it says the changes do not remove or diminish the underlying requirements within PCI DSS. FAQ 1588 says that criterion only applies to e-commerce merchants whose webpage includes a provider's embedded payment page or form, such as an iframe, and that the merchant can confirm it either with techniques such as, but not limited to, those detailed in 6.4.3 and 11.6.1, or with a confirmation from its PCI DSS compliant provider that the provider's solution, when implemented according to the provider's instructions, includes techniques that protect the merchant's payment page from script attacks. Ask your acquirer or payment brand whether SAQ A is the questionnaire for you.
FAQ 1588 says it does not apply to e-commerce merchants with a webpage that redirects customers from the merchant's webpage to the provider, with examples including an HTTP 30x redirect, a meta redirect tag or a JavaScript redirect, or to merchants that fully outsource payment, for example by emailing customers a link to the provider's website. That answer is about one eligibility criterion only, and the FAQ says its contents should not be interpreted to impact or contradict any other eligibility criteria in SAQ A or in any other SAQ. The same FAQ says to consult the entity you would submit a questionnaire to, typically an acquirer or the payment brands, to determine whether you must submit one and which. Record the redirect in rows L1 and A1 and put the question to them in A2.
For 3DS, FAQ 1581 says that for merchants using a 3DS solution, validation to 6.4.3 for 3DS scripts is not required, due to the inherent trust relationship established through the merchant's due diligence, onboarding and business agreement with the 3DS service provider, and that any script run outside the purpose of performing 3DS functionality is subject to 6.4.3. FAQ 1592 deals with Requirements 12.8 and 12.9 instead: as part of an entity's assessment of its e-commerce environment, a provider of third-party scripts is not considered a third-party service provider for those requirements if the entity confirms both that the provider's only service is providing scripts not related to payment processing and that those scripts cannot impact the security of cardholder data and/or sensitive authentication data. List every script in the ledger with its purpose, and put any question about a particular one to your acquirer or assessor.
No. A review or test records how a payment page behaved on the build and day it covered. The Council's post of 30 January 2025 says compliance validation requirements are set by compliance enforcing entities such as brands and acquirers, and FAQ 1588 has the merchant itself confirm the SAQ A script criterion. Clever Ops is neither an Approved Scanning Vendor nor a Qualified Security Assessor. Its reports are not certificates: they give the merchant, its acquirer and any assessor dated evidence to weigh.
Scope the right test for your application
Tell us what you have built, who uses it and when it needs to launch. You will get a plain-English recommendation: a website audit, a web application test or a referral to a specialist.