Unfortunately, nothing like CF exists in math because rating mathematical proofs is quite difficult, since unlike code, math proofs are not always built in some unified formal language. If you just want to improve your number theory skills it seems that solving past school olympiads like BMO/IZhO/RMM etc or books like "NUMBER THEORY Structures, Examples, and Problems" by Titu Andreescu is better, Well it is better if you look for smth similar to cp problems, if you looking for some uni level research problem then idk
Unfortunately, nothing like CF exists in math because rating mathematical proofs is quite difficult, since unlike code, math proofs are not always built in some unified formal language. If you just want to improve your number theory skills it seems that solving past school olympiads like BMO/IZhO/RMM etc or books like "NUMBER THEORY Structures, Examples, and Problems" by Titu Andreescu is better, Well it is better if you look for smth similar to cp problems, if you looking for some uni level research problem then idk