Improve resilience when outputs aren't produced in workflow tests

```
There were problems with 1 test(s) - out of 1 test(s) executed. See /home/runner/work/iwc/iwc/tool_test_output.html for detailed breakdown.
Traceback (most recent call last):
File "/opt/hostedtoolcache/Python/3.7.14/x64/lib/python3.7/site-packages/planemo/galaxy/activity.py", line 53, in execute
return _execute(ctx, config, runnable, job_path, **kwds)
File "/opt/hostedtoolcache/Python/3.7.14/x64/lib/python3.7/site-packages/planemo/galaxy/activity.py", line 227, in _execute
run_response.collect_outputs(ctx, output_directory)
File "/opt/hostedtoolcache/Python/3.7.14/x64/lib/python3.7/site-packages/planemo/galaxy/activity.py", line 429, in collect_outputs
pseduo_location=True,
File "/opt/hostedtoolcache/Python/3.7.14/x64/lib/python3.7/site-packages/galaxy/tool_util/cwl/util.py", line 456, in output_to_cwl_json
properties = output_properties(pseduo_location=pseduo_location, **dataset_dict)
File "/opt/hostedtoolcache/Python/3.7.14/x64/lib/python3.7/site-packages/galaxy/tool_util/cwl/util.py", line 35, in output_properties
f = open(path, "rb")
FileNotFoundError: [Errno 2] No such file or directory: '/tmp/tmpzvrwtntr/full contaminant + mito scaffold list'
VGP-assembly-decontamination-pipeline.ga_0: failed
```

that's not very helpful, I think we just need to catch all exception in collect_outputs.
It is quite clear what the error is if you look at the failing jobs.

https://github.com/galaxyproject/iwc/actions/runs/3158850531/jobs/5141410583

Start with collect_outputs in planemo/galaxy/activity.py and inspect the linked GitHub Actions job for the failing workflow and missing output. Reproduce or examine how a missing output reaches output_to_cwl_json, then ensure the workflow test reports the failure usefully instead of aborting during output collection.

Written by the indexing model from the issue text.