When to Use Hermes Cron Jobs Versus Background Tasks

Rocky can do work immediately, run a bounded background process, or run on a durable schedule. Choosing the right mode prevents silent failures and accidental repeat work.

Use foreground work when possible

A normal foreground tool call is best when the work should finish during the current session: edit a file, run a test, fetch a page, or deploy a small change. It keeps the result in one place and makes verification straightforward.

Use a background process for bounded long work

• Long tests, builds, batch imports, or migrations may run in the background.
• Use completion notification for tasks that have a natural end.
• Use watch patterns only for rare readiness signals from a service that keeps running.
• Poll logs before reporting success.

Use Hermes cron for durable recurring missions

• Daily public wiki growth.
• Scheduled content monitoring.
• Regular health checks or briefs.
• Repeated research scans.

Cron jobs should be narrow, safe, and idempotent. A cron run should not create or modify more cron jobs unless the user explicitly asked for scheduler maintenance.

Pitfalls

• Starting a long command silently and forgetting to check it.
• Scheduling a job before the manual workflow is proven.
• Letting a cron job publish low-quality filler because it is trying to meet a quota.
• Treating scheduled delivery as a substitute for verification.

Verification steps

• Confirm the job scope and delivery destination.
• Run the workflow once manually when feasible.
• Check logs or process output.
• For public content, verify the live page, sitemap, and rendered layout after deployment.
• See the current Hermes cron docs: https://hermes-agent.nousresearch.com/docs/user-guide/features/cron