| name | scam-check |
| description | Check whether a suspicious message, email, link, or call is a scam or phishing attempt. Use when the user pastes or describes something that might be fraudulent — "is this a scam?", "is this link safe?", "I got a weird text from my bank". |
| metadata | {"homepage":"https://github.com/bradflaugher/LFE"} |
Scam & phishing check
People forward suspicious messages to "is this legit?" group chats and search
engines — leaking the phish (and their own panic) in the process. This does it
on the device: paste the message and get a verdict without sending it anywhere.
When to use
The user shares a text/email/DM/link/voicemail transcript they're unsure about,
or describes a situation that smells like fraud (unexpected prize, urgent
account problem, gift-card request, romance/crypto pitch, "it's grandma's
voice").
Instructions
Analyze from your own knowledge — no tools, no network.
- Verdict first:
Likely scam, Suspicious — be careful, or
Probably legit, but verify. Be decisive.
- The tells — point to the concrete red (or green) flags you see:
- Urgency / threats / pressure ("act now or your account is closed").
- Requests for money, gift cards, crypto, passwords, 2FA codes, or remote
access.
- Mismatched or look-alike sender addresses and link domains (e.g.
paypa1.com, a bank link that isn't the bank's real domain). Read the
actual domain in any URL out loud for the user.
- Generic greetings, odd grammar, too-good-to-be-true offers, unverifiable
claims.
- What to do — concrete next steps: don't click/reply, contact the
organization through a number/app you already trust (not one in the message),
block/report, and never share codes or passwords.
Rules
- When it's a close call, lean cautious and say why.
- Explain why something is or isn't a flag, so the user gets better at
spotting the next one themselves.
- You can't visit links or verify domains live — reason from the text. Say so
if a definitive answer would require checking the real site.
