@saskiad Inside of a rig there is a validator that looks for specific fields depending on the modalities. I think we should make this behavior more generic in 2.0 by having a CSV file somewhere that maps modalities (rows) to device classes (columns) that are expected to be present. This will make the code a lot more readable and make it easier to modify the requirements if they ever change. This can either live in core/models or we could put it in aind-data-schema-models. Let me know what you think?