The Idiot Files: How I Let My X Account Get Hacked

A DM:

Message congratulating Chris on the Codex AI launch and suggesting collaboration between OpenAI and WebMCP on agent-driven development workflows.

Well, thank you! We really did recently launch CodePen AI and WebMCP Support.

Who is this person I don’t recognize?

Social media profile for Tricia Troth, Head GTM at OpenAI, showing her portrait, London location, and follower information.

Tricia Troth. Never met her. But she’s at OpenAI, eh? And a decent number of followers. That tracks. I feel like there are people like Tricia at companies like OpenAI that aren’t that into social media, but the fact that they work there gets them a bunch of followers. Felt like it checked out. Verified account as well.

There was a 🚩 red flag, though.

Screenshot of a direct message conversation showing a partnership discussion about CodePen AI and a Calendly scheduling link.

“Username changed recently”, which apparently X flags as potentially bad news. Apparently that wasn’t entirely enough to turn me away, though.

“Tricia” even has a believable LinkedIn account.

LinkedIn profile page for Tricia Troth, Head of GTM, Startups EMEA at OpenAI, with her portrait and connection details.

Calendly is a real, normal website that people use to book meetings. The language felt normal and natural. Of course, I’d love to take that meeting.

Clicking over to that site, it really was calendly.com. Not a fake site. I picked a time that worked for me.

Then it redirected back to X to approve some app. Here’s the app I screenshot as I was removing it after I was hacked:

Multi-Platform Sync app permissions page showing access to view posts, upload media, and follow or unfollow accounts.

This should have been another 🚩 red flag. Those are way too many permissions for some kind of weird calendar flow thing.

What my brain did was: OK this is some friggin AI app thingy. There are probably all sorts of wild AI mandates at OpenAI, where Tricia vibe-coded some kind of personal CRM to keep track of her leads, so she can explain to her boss how she’s all AI’d up or whatever.

But of course, approving this app is what “got hacked” means here. Now whoever controls this app can post on my behalf.

And post they did. Fortunately I have actual friends who let me know right away.

Screenshot of a message sharing an X post introducing Rulebook, a tool for testing customer support bots.

I was able to revoke the app access quickly, but not before a couple of more much worse posts. I deleted them, and that was that.

There was some fallout from some other horrible accounts accusing me of being a scammer or like pumping a memecoin or something? Which wasn’t even what the hacked posts were about? Just gross crap.

Don’t worry everyone, I’ve reported the user, and X will definitely take care of it!!

X Support notification stating that a reported account did not violate rules against illegal and regulated behaviors.

Hahahhaha just kidding, X doesn’t give two fucks.

They are too busy making sure to remove my old harmless jokes.

Screenshot of an X content privacy notice addressed to Chris Coyier, displaying a post graphic about tabs versus spaces and coding standards.

Leave a Reply

Your email address will not be published. Required fields are marked *