lightscanner
Privacy compliance POC that crawls websites with Global Privacy Control enabled to detect unauthorized third-party trackers.
3 commits over 2 days
What it is
LightScanner simulates a browser session with the Global Privacy Control (GPC) signal enabled and tests whether websites respect the user's "do not track" request. It loads pages using stealth browser automation, captures all network requests, filters out first-party traffic, and flags any known third-party trackers that violate privacy expectations. The tool is designed for enterprises and privacy advocates who need an objective, automatable way to validate real-world GDPR/CCPA compliance.
Why I built it
Built to validate whether websites claiming privacy compliance actually honor GPC signals in practice. The PRD indicates this was scoped and commissioned by "Sentinel Insights — Fractional CTO Team," suggesting it was a client deliverable or proof-of-concept project to demonstrate automated privacy-compliance auditing capabilities for enterprise customers.
Tech stack
Node.js with TypeScript, Puppeteer-extra with stealth plugin for anti-detection browser automation, Jest for testing, and custom HTML/JSON report generators. Uses tracker blocklists (EasyPrivacy/Disconnect-style) stored in JSON config files.
Timeline & effort
Last active Jun 20, 2025 — 3 commits across 2 active days (spanning 0 days).
Work spanned two active days (June 19–20, 2025) with just 3 commits. The commit data is sparse—likely bundled progress or work tracked elsewhere—but the tasklist shows substantial completion across seven phases (project setup, browser automation, request monitoring, tracker detection, reporting, and testing). The codebase includes a modular TypeScript architecture, multiple reporters (console, JSON, HTML), unit tests, and several generated reports, indicating meaningful development effort despite minimal commit history.
Did it turn out useful?
The PRD is marked "Draft — for implementation in Cursor" and the tasklist shows many core features checked off but several polish/advanced tasks incomplete (confidence scoring, progress indicators, CI/CD, comprehensive docs, concurrent scanning). Generated HTML reports in the repo (bestbuy, cnn, foxnews, techcrunch, parallax) show the scanner has been run successfully, proving the core concept works. However, with only 3 commits, private visibility, zero stars, and no deploy URL, this looks like a proof-of-concept delivered to a client ("Sentinel Insights") rather than a shipped product or ongoing tool.