Create a tool to help debugging failed runs

[wonder-sk]

Sometimes when a work-packages run fails with some data issue, it would be useful to have a tool to recreate the case by gathering all the input data at a particular date/time, so we can reliably run the case again and figure out the issue.