You're making an assumption here that hasn't been confirmed. That assumption being that any app accessing user data from Meta is proxying (i.e., streaming the requested data to the app publisher's servers and then passing that data along to the end user) that data through their servers.
Is that the case with the app in question? Is it the case with every such app?
Or are there apps that directly connect to Meta's servers from the user's hardware without streaming the requested data through the app publisher's servers?
The app in TFA may be proxying (see above) data through their servers (that's the definition of a proxy in this context), but I don't know if they are doing so. If they are, there certainly are serious privacy/security issues with that process.
But again, no one has provided evidence that's what the app in question is doing. If they are, you should run screaming in the other direction.
However, if the app is simply performing the same API calls as Meta's app and returning the data directly to the end user, the risk profile is pretty similar for both apps (dependent on code quality, the ethical stance of the publishers, etc.).
If you use the built-in dev tools to do the same thing, then there is no proxy.
An alternative client for something is (usually) not a proxy. It connects directly.
But more importantly, "a proxy where the app developers can intercept and see all data" is not referring to a client-side proxy. Even if there was a client-side proxy involved somewhere, that would make the initial claim wrong.
The app is by definition acting as a proxy, and therefore the developers can intercept and see the data, though they might not be doing so currently.