Hashing a MD5 digest with a private key, then XOR[ing] the digest and private key [migrated]

TL;DR
Does hashing a MD5 digest with a private key, then XOR[ing] the digest (twice recursively, with offset) have any glaring issues? Would this mean MD5 is predictable?
Is partial MD5 recognition possible?
General Overview
I convinced my… Continue reading Hashing a MD5 digest with a private key, then XOR[ing] the digest and private key [migrated]