If I am given a problem in P, to claim that I have solved it, I would need to give proof, right? For example, to show a number is not prime, if I use some deterministic algorithm, I can return the prime factors of that number.
I am finding it weird that P does not require a witness/certificate in the definition like NP.