I see that you've found the import:dip-objects command
. That was really the only suggestion I had. To clarify theCLI task requirements and behaviors:
If, as in the example shown in the documentation, you provide the same target slug value for every DIP object in the CSV you create, then the upload behavior of the CLI task will be identical to a basic DIP upload from Archivematica - all objects will be attached to new stub descriptions as children of the target description, and their slugs and description names will come from the object. The CSV merely gives you an option to send different DIP objects to different descriptions, something only possible in Archivematica if you are using the Appraisal tab for arrangement.
However, if you don't want to manually create the CSV or the task behavior won't meet your needs, then my best suggestion would be to use Archivematica's Partial re-ingest
to generate a new DIP and upload it to AtoM that way.