How a site audit works
This lab explains the audit. It does not fetch a URL. The scanner is on Black Box Site Design. If that page is not live yet, the address below is still where it will be.
The run stops after 30 seconds, loads 1 page, and allows 3 audits an hour for one address.
- One public page is opened in a sandboxed headless browser. The run is a GET and a render. It does not submit a form, click a login, or run a script the visitor typed.
- Only http and https are allowed. DNS has to resolve to a public address. Localhost, link-local, and private ranges are refused, and a redirect is checked again.
- One address can run three audits in an hour, and there is a daily cap for everyone. The address is used for that cap and is not written to a log. The page body is not stored.
- The result names the tags, the dataLayer event names and keys, the cookies set before consent, the lab vitals, the missing basics, and the broken scripts.
- An MCP tool named audit_page uses the same limits. This origin does not expose that tool. The scanner is the Site Design audit.
Understanding the labHow to use it and why you'd want to
Why you'd want to
You want a page audit that cannot reach a private network. This lab names the sandbox, the checks, and the rate limit, then points at the scanner.
How to use it
- Read the sandbox, the address checks, and the cap.
- Open the Site Design audit. It may not be live yet.
- This page does not fetch a URL.
WalkthroughHow we built itShowHide
Example vendor setup is marked below. eVar and event numbers are examples. This lab does not call Adobe.
The site audit lab explains a sandboxed browser, address checks, and a rate limit. The scanner runs on Site Design. This page does not call Adobe.
Name the sandbox
One public page is rendered. The run does not submit a form or click a login.
Refuse private addresses
Only http and https pass, and DNS has to resolve to a public address. A redirect is checked again.
Cap the runs
Three audits an hour for one address, a daily cap, and a 30 second stop. The address is not logged.
Point at the scanner
audit_page would use the same limits. This origin does not expose it. The link goes to Site Design.
From the code
The public address of the scanner, and the limits this lab names.
export const SITE_AUDIT_URL = "https://www.blackboxsitedesign.com/audit"
export const auditLimits = {
timeoutSeconds: 30,
pages: 1,
perAddressPerHour: 3,The browser rule and the address rule. The page prints these strings.
en: "One public page is opened in a sandboxed headless browser. The run is a GET and a render. It does not submit a form, click a login, or run a script the visitor typed.",
es: "Se abre una página pública en un navegador headless aislado. La ejecución es un GET y un render. No envía un formulario, no pulsa un inicio de sesión y no corre un script escrito por el visitante.",
},
{
en: "Only http and https are allowed. DNS has to resolve to a public address. Localhost, link-local, and private ranges are refused, and a redirect is checked again.",Google and Adobe
This lab does not queue its own event. The Google block shows an example of how you would track the interaction. Adobe is an example too.
This lab does not queue an analytics event of its own. After Accept, the site still queues virtual_page_view for the page, with page_path and page_title. Clicks inside a data-track-ignore region are not recorded. The push below is an example, not a call this page makes. This site does not call gtag('set') or Adobe.
Example only. This page does not push this object.
// Example only. This lab does not queue this event.
dataLayer.push({
event: "site_audit_example",
label: "architecture",
})Example implementation. This site does not call Adobe. eVar124, prop124, and event124 are example variable numbers. In Launch, a data element reads the value and a rule sends it when the value changes. In the report suite, enable eVar124 with Most Recent allocation and expire it on the hit. A processing rule is unnecessary if the data element maps straight to that eVar.
Example alloy payload. _example is a placeholder schema, not a real tenant.
alloy("sendEvent", {
xdm: {
eventType: "web.webinteraction.linkClicks",
_example: {
auditLimit: "one-page",
},
},
})Example hit. eVar124 is hit-scoped in this example. prop124 dies at the same hit.
s.eVar124 = "one-page"
s.prop124 = "one-page"
s.events = "event124"
s.linkTrackVars = "eVar124,prop124,events"
s.linkTrackEvents = "event124"
s.tl(true, "o", "site_audit_example")Translation
Each lab field, the GA4 name, and the example Adobe variable. Scope is the part that usually gets lost in the translation.
Page limit
GA4. Not sent. This lab does not queue an event.
Adobe. Example event124. This site does not call Adobe.
Scope. A limit on the run, not a hit.
Address check
GA4. Not sent.
Adobe. Example eVar124. This site does not call Adobe.
Scope. Checked before the request.
Tool name
GA4. Not sent. audit_page is described, not called.
Adobe. Do not put the tool call in a prop. Not sent.
Scope. It stays in the explanation.
Consent
GA4. Not a parameter. Opening the lab still follows the site page view.
Adobe. Example: do not add a hit for reading this page.
Scope. The site page view waits for Accept.
A tag audit for your own site is a conversation, not a button on this page. Contact
Recommended next
Why these picksReading path transitions…