Since there's no ether transfer between the contracts, there鈥檚 not much room from attack from a Force Sending perspective. There are ways where one can track each of the owners of a split secret based on the transactions, but since everything was previously encrypted end-to-end, there's not much use with that information, which is even public on an IPFS Hash.
The biggest consideration was considering the fixed size of values for hashes. Although string could had been used, we preferred to stick to bytes32 to avoid any over-use of memory or any memory related attacks.