Foundry should use (or know about) job cache

This means
- adding `--use_cache` on the planemo side (planemo rerun has it, planemo run does not)
- teaching foundry to use a project specific profile with a persistent database
- teaching foundry about failure modes (green but crashed tools ... should be very rare)

Start by comparing how planemo rerun handles --use_cache with planemo run, then inspect Foundry's project profile and persistent database setup. Review the existing failure handling for crashed tools. Done means cache use is supported on the planemo side, Foundry uses a project-specific persistent profile, and relevant failure modes are handled.

Written by the indexing model from the issue text.