This seems like a good approach, but you can still get around it by detecting apple reviews and not doing anything with private APIs during them
[1] https://developer.apple.com/library/ios/documentation/Genera...
There's even a paper on intentionally inserting security flaws into your code and then exploiting them from your own server to change execution patterns:
https://www.usenix.org/conference/usenixsecurity13/technical...
Ultimately, you need to enforce access control instead of just trying to detect problems a priori. Apple's sandbox is a great start to that, and I expect they'll keep improving it to block apps like these.