checkbox 7.0.0 on 25.10 traceback when running subsequent test runs after the first test run completes.

### Bug Description

I am doing 25.10 validation on Z/VM and Z/LPAR instances as part of the deliverables to IBM each release cycle.

I installed checkbox 7.0.0 from the beta PPA on these machines and ran the full cert run successfully, and at the end submitted the results.

The results indicated some issues with LXD testing so I tried re-running the virt tests using the test-virtualization launcher.

Instead of the tests running, I got a traceback (below).

I had to add --clear-cache and --clear-old-sessions in order to successfully run subsequent tests. (clear-cache is probably not needed, but I added both as a "Nuke it from Orbit" option.

I presume this will happen on all architectures, but I have only tried this on s390x so far

### Cert-blocker Test Case

- [ ] cert-blocker

### To Reproduce

Install checkbox (canonical-certification-server) from beta PPA (presumably on any arch, I did this on s390x)
run a test and submit results at the end via checkbox (I ran the cert suite for s390 validation)
run a retest (for example test-virtualization)
Checkbox starts and dumps the traceback.

### Expected Result

I can run launchers without issue after the first run is complete

### Actual Result

traceback when trying to run a subsequent test / launcher.

I can run test runs after the initial one only if I also add --clear-cache and --clear-old-sessions (at oeast one of these, likely clear-old-sessions resolves whatever causes the traceback).

### Environment

Ubuntu 25.10 on IBM z15 s390x
Checkbox 7.0.0.dev26

### Relevant log output

```shell
ubuntu@hwe0008:~$ test-virtualization
WARNING:plainbox.secure.providers.v1:Skipped file: /usr/share/checkbox-provider-base/units/camera/README.rst
WARNING:plainbox.ctrl:warning: /usr/share/checkbox-provider-base/units/suspend/suspend.pxu:657-673: job 'suspend/iperf_after_suspend_ether_auto_device1_enc600', field 'summary', please stay under 80 characters
WARNING:plainbox.ctrl:warning: /usr/share/checkbox-provider-base/units/suspend/suspend.pxu:657-673: job 'suspend/iperf_after_suspend_ether_auto_device1_enc600', field 'summary', please stay under 80 characters
Traceback (most recent call last):
File "/usr/bin/checkbox-cli", line 8, in
sys.exit(main())
~~~~^^
File "/usr/lib/python3/dist-packages/checkbox_ng/launcher/checkbox_cli.py", line 183, in main
return subcmd.invoked(ctx)
~~~~~~~~~~~~~~^^^^^
File "/usr/lib/python3/dist-packages/checkbox_ng/launcher/subcommands.py", line 262, in invoked
if not self._auto_resume_session(self.resume_candidates):
~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/checkbox_ng/launcher/subcommands.py", line 436, in _auto_resume_session
self._resume_session(last_session.id, None)
~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/checkbox_ng/launcher/subcommands.py", line 615, in _resume_session
raise ValueError(
"Unsupported outcome for resume {}".format(outcome)
)
ValueError: Unsupported outcome for resume not-supported

ubuntu@hwe0008:~$ checkbox-cli --version
7.0.0.dev26
```

### Additional context

_No response_

Start by reproducing a second launcher run after a completed and submitted run, first without and then with --clear-old-sessions. Read checkbox_ng/launcher/subcommands.py around _auto_resume_session and _resume_session, following the call from checkbox_cli.py and the reported “Unsupported outcome for resume not-supported” traceback. Done means subsequent launchers run without requiring cache or session-clearing flags.

Written by the indexing model from the issue text.