Nexus: missing executable not caught in a ws16 run

**Describe the bug**
Follow-on to #6203 with a missing pw.x

After fixing the oversubscription
```
$ cat ~/.prte/mca-params.conf
rmaps_default_mapping_policy = :oversubscribe
hwloc_base_binding_policy = none
```

We now get the same output from the Nexus workflow but:

```
$ cat runs/diamond/scf/scf.err
--------------------------------------------------------------------------
prterun was unable to find the specified executable file, and therefore did
not launch the job. This error was first reported for process rank
0; it may have occurred for other processes as well.

NOTE: A common cause for this error is misspelling a prterun command
line parameter option (remember that prterun interprets the first
unrecognized command line token as the executable).

Node: mymachine
Executable: pw.x
--------------------------------------------------------------------------
```
This is the correct error.

Again, scf.out is zero length, as expected.

Hopefully a one line update to catch "unable to find" or the like is enough.

This would be an easy problem for someone new to hit since it is easy to miss fully setting up the environment to make all required simulation codes available on the PATH.

**To Reproduce**
As #6203

**Expected behavior**
Catch the error promptly.

**System:**
macOS laptop, macports dependencies

Reproduce the Nexus workflow described in #6203 with pw.x absent from PATH, then inspect the workflow entry point that handles prterun failures and the resulting runs/diamond/scf/scf.err. Identify how the "unable to find" executable message is surfaced; done means the missing pw.x is reported promptly instead of allowing the workflow to continue with an empty scf.out.

Written by the indexing model from the issue text.