Hooks are not always checks, shall we update the hooks interface to reflect this?

[auryn-macmillan]

The interface for hooks describes the pre and post execution calls as "checks", preCheck and postCheck.

This is descriptive of how guards have been used historically and how many hooks will presumably be used in future. However, it's perhaps too limiting of a descriptor as some use-cases will not be checks, per se, but additional code that must be executed alongside each transaction. This is presumably the rationale for changing the name from "guard" to "hook", so it stands to reason that the rest of the interface should reflect this as well.

I'd propose something like the following changes.

interface ISafeProtocolHooks {
    function preExecutionHook(Safe safe, SafeTransaction tx, uint256 executionType, bytes calldata executionMeta) external returns (bytes memory preCheckData);

    function preRootAccessHook(Safe safe, SafeRootAccess rootAccess, uint256 executionType, bytes calldata executionMeta) external returns (bytes memory preCheckData);

    function postExecutionHook(Safe safe, bool success, bytes calldata preCheckData) external;
}