`complete` .ind files in task plans could store data hash

I think they could, anyway. I'm looking at a preliminary task remake file from the lake temperature stuff, which has tasks like this:
```
# --- NLDAS_7.12_223.223_141.141_dlwrfsfc.nc --- #

1_nldas_cubes/out/NLDAS_7.12_223.223_141.141_dlwrfsfc.nc:
command: nccopy_nldas(target_name)

ind_dir/NLDAS_7.12_223.223_141.141_dlwrfsfc.nc.ind:
command: sc_indicate(target_name)
depends:
- 1_nldas_cubes/out/NLDAS_7.12_223.223_141.141_dlwrfsfc.nc
```
(see https://github.com/USGS-R/lake-temperature-model-prep/pull/3/files?utf8=%E2%9C%93&diff=split#diff-989a07de296dcf114a762d038357bdf0)

I actually don't think it'll change behavior much, but could be helpful for interpretation of rebuilds, if we included hashes of the depends files within the .ind file created by sc_indicate.

Start by reading the sc_indicate entry point and the task-plan examples in the issue to understand how complete .ind files are created. Determine the expected representation of hashes for dependency files and define how the resulting metadata would support interpreting rebuilds; the issue is done when that behavior and its validation are specified.

Written by the indexing model from the issue text.