`mechababs jobs` needs a `has_results` column

## What

babs's `job_status.csv` carries `has_results`, its real did-this-produce-output signal, and `mechababs jobs` does not render it.

Symptom that surfaced it (2026-09-01, real run): a **merged** cell's job renders with a blank `state` (and blank `time_used`/`time_limit`) while `sacct` still reports it `COMPLETED 0:0` — so the one row where the drill-down should be most conclusive is the one that goes quiet.
`has_results` is the column that stays true after the scheduler forgets the job, which is exactly the post-merge case.

## Fix

Add the column to the `jobs` table, sourced from babs's csv like the others.
Additive, no correctness change.

Start at the implementation of the `mechababs jobs` table and trace how its existing columns are sourced from babs's `job_status.csv`. Add `has_results` alongside those columns, then run `mechababs jobs` against a representative job-status file and confirm the new value is rendered.

Written by the indexing model from the issue text.