When someone receives a suspicious bank fraud text, a WhatsApp remote-job offer, or a QR code parking notice, their biggest hesitation before using an online scam checker is privacy: "If I paste my personal SMS or email here, is a server logging my phone number and banking details?"

That question is why we architected ScamShield Lab (https://scamshieldlab.com) around a strict zero-storage, client-side execution model.

1. The "Airplane-Mode" Privacy Architecture

Instead of sending user messages to a backend API or database, ScamShield Lab executes its entire 4-layer heuristic scoring engine locally in browser volatile memory:

  • Layer 1 (Cognitive Urgency & Coercion): Detects artificial countdowns, arrest threats, and loss-aversion triggers.
  • Layer 2 (Domain & Optical Routing): Flags lookalike combo-squatting domains, Punycode (xn--), and reverse-proxy Evilginx/AiTM lures.
  • Layer 3 (Irreversible Payment Extraction): Identifies Zelle "safe holding account" traps, Regulation CC fake check float lures, and ERC-20 Permit wallet drainers.
  • Layer 4 (Out-of-Band Verification Mismatch): Evaluates whether the message violates standard institutional due-process rules.

Anyone can verify that zero bytes leave their device in 10 seconds: load the page, turn off Wi-Fi (or open DevTools → Network tab), and run a scan. It evaluates locally with zero XHR/Fetch calls. You can read the full scoring breakdown here:
https://scamshieldlab.com/how-we-score

2. Defending Against 2026's Fastest-Growing Vectors

Alongside the scanner, we published 18 open-access forensic guides and interactive household defense tools:

If you work in IAM, SOC operations, or consumer security awareness, I'd love your feedback on additional heuristic patterns we should add next!