Instant download
Hermes Operator Setup Kit
Use this to get Hermes into a stable operator shape before you add more workflows, cron jobs, or public-facing automation.
Setup checklist
Cron audit
Execution patterns
First wins
1. Stand up the operator baseline
- Pick one owner for Hermes runtime, profiles, and cron health.
- Document which model/provider each workflow should use before you scale job count.
- Separate low-risk internal jobs from anything client-facing or publish-facing.
- Write down where outputs land, who reviews them, and what “healthy” means.
2. First deployment checklist
- Confirm profile config, keys, and filesystem paths are explicit.
- Run one known-good test task per model route.
- Run one cron manually before trusting the schedule.
- Verify output location, not just process status.
- Record rollback notes for every live automation.
3. Cron audit worksheet
- What is the exact job output?
- Where is the output stored?
- What stale window is acceptable for this schedule?
- What makes the output suspicious or placeholder-grade?
- What is the smallest deterministic fallback if the agent fails?
4. Execution pattern rules
- Tell the worker the exact task, artifact, and finish condition.
- Prefer explicit verbs over broad requests like “handle this.”
- Make the output target concrete: file path, status update, or summary block.
- Separate research, execution, and publishing into different prompts when risk differs.
5. First visible wins
- One daily status report that actually reflects live state.
- One queue-prep automation that saves manual sorting time.
- One diagnostic routine that catches stale jobs before the owner does.
Upgrade path
If you need this translated into live workflows, routing, review gates, or client-facing automation, move into a Free Assessment.