All projects
PG
AdTech2026

PixelGuard

Governance scanner that scores a marketing pixel before it ever ships.

Built for: Ad operations governance

Open the live app Source
loading demo…

Runs entirely in your browser with the same rule categories as the deployed scanner. Try the sample tags or paste your own. Nothing you paste leaves this page. The runtime headless-browser audit is not included here.

The problem

Tags arrive from vendors as opaque snippets. Ad ops has minutes to approve them, and a bad one can leak PII in a URL, load a hidden iframe, or fingerprint users. Security and legal find out after it is live.

How the real system works

  1. 01The static scanner parses the snippet: it extracts Floodlight src, type, cat, ord and u-variables, then lists every domain, script, iframe, image and beacon it references.
  2. 02Each finding is matched against a rule set: eval and new Function, document.write, hidden iframes, redirects, cookie and storage access, fingerprinting calls, and PII-shaped parameters such as email or phone.
  3. 03Findings are weighted into a risk score and a verdict. Main Floodlight risk is separated from piggybacked, non-Google calls so the reviewer sees who is really being loaded.
  4. 04A companion runtime version fires the tag in headless Chrome and captures every network request, catching piggybacks that only exist once the ad server responds.

What you take away

What actually hides inside a tracking pixel, and which patterns should stop a tag at review rather than after launch.