hermes - 💡(How to fix) Fix fix(cron): mark_job_run() auto-delete leaves orphaned output dirs when repeat limit is reached [1 pull requests]

Bug

`mark_job_run()` removes a job from `jobs.json` when its repeat limit is reached (lines 737–741 in `cron/jobs.py`), but does not delete `OUTPUT_DIR/<job_id>`. Each run of the job writes output files via `save_job_output()`, which creates `~/.hermes/cron/output/<job_id>/`. When the job auto-deletes on limit, that directory and all its `.md` files are left behind permanently.

`remove_job()` gained the same cleanup in PR #21882 (merged today). `mark_job_run()` is the only other path that deletes jobs and was not updated.

Steps to Reproduce

1. Create a one-shot cron job: `repeat: 1`
2. Let it run once — `save_job_output()` writes `~/.hermes/cron/output/<job_id>/`
3. `mark_job_run()` hits the repeat limit, removes the job from `jobs.json`
4. `~/.hermes/cron/output/<job_id>/` is never deleted

Expected

Output directory is removed when the job is auto-deleted, matching `remove_job()`.

Actual

Directory accumulates indefinitely after job completes.

Root Cause

`cron/jobs.py` lines 737–741:

if times is not None and times > 0 and completed >= times:
    jobs.pop(i)
    save_jobs(jobs)
    return  # ← OUTPUT_DIR / job_id left behind

`remove_job()` (lines 700–703) has the cleanup; `mark_job_run()` does not.