workflows for using confidential data

[tngTUDOR]

Do we already start building workflows for using confidential data, or do we leave this for later?

Some initial collaborations with private partners would require the use of confidential data, for certain, but do we need to provide support for such right away?

Also, during development, we might want to keep some data and models confidential, until the produced results are robust enough to be published. Is this a constraint we wish to always respect, or go fully open on every single data point?