Windshock Lens

Private, on-device scam and phishing analysis for your browser. Chrome built-in Gemini Nano + deterministic security rules.

Users0

Rating--

Reviews0

Manifest versionV3

7-day growth--

7-day growth rate--

Preview

Windshock Lens Media preview

6 assets

Trend

30-day user trend

Review user movement over the last 30 days.

User Growth Over Time

Rating trend

30-day rating change

Track rating movement over time to see whether quality signals remain stable.

Not enough rating snapshots yet. More data will appear as new snapshots are collected.

Growth overview

Daily, weekly, and monthly growth

Compare 1-day, 7-day, and 30-day net growth and growth rate.

1-day growthFlat

0--

7-day growthNo data

----

30-day growthNo data

----

Technical snapshot

Version, languages, and crawl freshness

Review publication date, version, supported languages, and crawl timestamps.

Version0.2.3

ManifestV3

Size19.07MiB

Languages1English

Published--

Store updated--

Last crawled--

Overview

Product summary

Review the store description, core capabilities, and common use cases.

Windshock Lens triages suspicious links and pages directly inside Chrome — before you click, before the page steals credentials, before a malicious download lands on disk. It is built for the gray zone that Chrome Safe Browsing and standard endpoint security tools miss: zero-hour brand-impersonation pages on free hosting platforms (workers.dev, pages.dev, firebaseapp.com, vercel.app, …), AI-client lookalikes, fake software download pages, and ClickFix shell-payload tricks.

How it works

Windshock Lens combines four independent signal layers:

1. Browser-side page extraction — the DOM, forms, links, clipboard writes, and downloads triggered by the target page are collected without sending the page anywhere.

2. On-device Gemini Nano LLM — Chrome's built-in language model evaluates the extracted signals locally. Page content, URLs, and OCR text never reach an external LLM API.

3. Deterministic security rules — hard evidence (shell payload on clipboard, dangerous URI schemes, auto-downloads, phishing-kit fingerprints) yields a verdict without the LLM. Brand-to-domain mismatch overrides catch impersonation patterns the LLM misses.

Reviews