Veriguard — scam check 作者： Aleks Linde

Paste a suspicious text, link, email or phone number — or, on desktop, select it on any page and right-click — and get an instant verdict explaining what's wrong with it.

HOW IT'S DIFFERENT

The detection engine is built into the extension. Your message is scored on your own machine, by rules you can read, and it never travels anywhere. Turn off your internet connection and it still works.

That's unusual enough to be worth stating precisely: the extension makes exactly one network request, and it isn't about you. It downloads a list of known malicious websites on a timer so it can recognise them offline. The request carries no query and no body — every copy asks for the same list the same way. There is deliberately no "is this site dangerous?" lookup, because answering that would mean telling us which sites you're checking.

WHAT IT CHECKS

· Links — lookalike domains, suspicious top-level domains, tracking parameters, known malicious hosts
· Messages — urgency and threat patterns, payment demands, impersonation of banks and government services
· Senders — mismatched reply-to addresses, failed authentication
· Phone numbers — number ranges and formats used by scam operations

Every verdict shows its working: the rules that fired, what each one contributed, and a score that adds up to the number shown. You can check our arithmetic.

WHAT IT WON'T DO

It won't follow shortened links. Resolving one would tell the scammer's link shortener your IP address, so it says the destination is unchecked instead of quietly guessing.

It won't pretend a quiet result is a clean one. If detection for your region is limited, or a check couldn't run, the verdict says so.

It won't ask to read the pages you visit. No host permissions, no content scripts. The only text it sees is text you give it.

PERMISSIONS

Two, both minimal:
· contextMenus — adds the right-click entry (desktop; Firefox for Android has no extension context menu, so there the toolbar button is the way in)
· storage — remembers your region and caches the malicious-site list

REPORTING

On a suspicious verdict you can report the scam to the public database. The extension opens the report form on veriguard.app with the link or number filled in — it never submits anything itself. You review it and send it.

OPEN SOURCE

The detection rules are public, because obscuring a keyword list wouldn't stop a sophisticated scammer — it would only stop you checking our work. The extension ships unminified so you can read what you installed.

github.com/alekslinde/veriguard

REGIONS

Scam patterns are local — the brands impersonated, the number formats, the agencies named. Veriguard ships rule packs for 27 countries across Europe, Asia, Africa, the Americas and Oceania, with the deepest coverage for Australia.

Anywhere else still works: the universal checks — lookalike domains, shorteners, tracking parameters, known malicious hosts, requests for personal details — run everywhere, and the verdict tells you when regional rules for your country aren't there yet rather than quietly scoring you as safe.