I hacked a dating app (and how not to treat a security researcher)

Returning the OTP in the request API response is wild. Like why?

I too am bewildered.

Maybe to make it easier to build the form accepting the OTP? Oversight?

I can't think of any other reasons.

Eliminate your database costs with this one easy trick!

One reason I could think of is that they may return the database (or cache, or something else) response after generating and storing the OTP. Quick POCs/MVPs often use their storage models for API responses to save time, and then it is an easy oversight...

Save a HTTP request, and faster UX! What's not to love?

When Pinterest's new API was released, they were spewing out everything about a user to any app using their OAuth integration, including their 2FA secrets. We reported and got a bounty, but this sort of shit winds up in big companies' APIs, who really should know better.

Oversight. Frameworks tend to make it easy to make an API endpoint by casting your model to JSON or something, but it's easy to forget you need to make specific fields hidden.

My best guess would be some form of testing before they added sending the "sending a message" part to the API. Build the OTP logic, the scaffolding... and add a way to make sure it returns what you expect. But yes absolutely wild.

So the UI can check if what they enter is correct.

It’s very sensible and an obvious solution if you don’t think about the security of it.

A dating app is one of the most dangerous kinds of app to make due to all the necessary PII. this is horrible.

> if you don’t think about the security of it.

This is big brain energy. Why bother needing to make yet another round trip request when you can just defer that nonsense to the client!

Smacks of vibe coding

I don't think a language model is that stupid. This smacks of pure human stupidity and/or offshoring.

I assume that whoever wrote it just has absolutely no mental model of security, has never been on the attacking side or realised that clients can't be trusted, and only implemented the OTP authentication because they were "going through the motions" that they'd seen other people implement.

No one would ever hack my app!

But LLMs are that stupid. Do you remember that guy who vibe coded a cheating tool for interviews and who literally leaked all his api keys/secrets to GitHub because neither him nor a LLM didn't know better?

Fair enough. Since it's trained on human stupidity, I suppose it would reflect that stupidity as well.

I’ve seen banks where the OTP code is generated on the client and then sent to the server.

Is that the same guy who had his degree revoked for creating a cheating tool for interviews and is now a millionaire for creating a cheating tool for interviews?

Everyone that programs should take blackhat classes of some kind. I talk to so many programmers that really don't understand what hackers/attackers can actually do.

Could be. Somewhere else in these comments someone was saying they found evidence that the app was coded that way.

But they also said it was a project by two students. And I could absolutely see students (or even normal developers) who aren’t used to thinking about security make that mistake. It is a very obvious way to implement it.

In retrospect I know that my senior project had some giant security issues. There were more things to look out for than I knew about at that time.

It appears that the OTP is sent from "the response from triggering the one-time password".

I suspect it's a framework thing; they're probably directly serializing an object that's put in the database (ORM or other storage system) to what's returned via HTTP.

that's my first thought at as well - like a basic CRUD operation that returns the row that was created as a response.