Possibly the most entertaining thing that you can do with AI: scamming the scammers.
You know when a "recruiter" messages you with a dream job, i.e. 150k a year for 5 to 10 hours a month, senior advisor role, at a fairly well known company?
Normally you just block it and move on. However, what you can do instead is hand the whole thing to Claude and ask it to find out who was behind it.
In about twenty minutes, here is what it dug up and did, most of it on its own while I was busy with other things:
- looked up the domain records and saw it was bought three days ago, with a fake name and address, but a real email and phone number left behind
- found the web host, the backend server they were hiding, and the email service they were using to collect whatever victims typed in
- spotted that they had put a real company's name into the site's code, to look trustworthy
- wrote up the abuse reports and had them ready for the host, the registrar, the backend provider and the browser blocklists
- drafted a message to send to the scammer
Thankfully, all of this is public information and anyone can look it up if they know where to look. The only thing that changed is that the digging I would have done by hand over an afternoon mostly happened in a couple minutes without my intervention. And honestly I would usually just block the scammer.
I cleaned the whole process up into a reusable skill, so I don't start from scratch next time and you can freely use it too. It is on GitHub and you can run it yourself in Claude or Claude Code.
So if someone tries to scam you, you don't have to just block and forget. You can take their whole setup down instead.