This is the approach I take.
I have a single Job and list my Object names in tFixedFlowInput, or I could externalise them.
The Job iterates through these, extracts the Salesforce Ids to a file and then uses that file to perform a bulk delete.
This approach also allows me to constrain the deletes based on Owner, Creation Timestamp and Record Id etc.
It also allows me to place a tMsgBox between the Id collection and delete (showing some useful information) just incase I'm pointing to the wrong Org!
This is a simple Job that is simply driven by a list of Object names, and using components that are configured by these.