Define benchcab workflow for model output

[abhaasgoyal]

With v3.0.0 of me.org API, there is support for creating model output endpoints, and adding corresponding experiments/benchmarks.

Requirements

• Use benchcab with hardcoded values for supported experiments/benchmarks, so that the user does not need to provide a fixed model output ID.
• Derive model output name from branch/local path (query user if multiple paths exist)
• Determine when to delete_all files