What problem does it solve? A rebuilt Hermes agent loses its cron schedule because nothing replays jobs.json, leaving the life dashboard's wall screen permanently stale with no way to diff what was lost. This Skill replays a reviewed, versioned spec of all seven producer schedules instead of improvising them from memory. ## Core Features & Use Cases - Idempotent cron registration: Reads hermes cron's own jobs.json state file and creates only missing jobs, so it is safe to re-run after any rebuild. - Defensive validation: Refuses to register when the container timezone disagrees with family.timezone, when the Plow Chat delivery target is unset, or when the jobs file is unreadable — never mistaking "can't tell" for "nothing is registered". - Paused-job detection: Names producers that are registered but paused and exits non-zero so unattended re-provisioning can act on the signal. - Use Case: After rebuilding the agent's home volume, ask the agent to set up the life dashboard crons; the script registers all six producers (weather, sports, morning updates, morning/evening triage, weekly digest, calendar nudge) and reports exactly what happened. ## Quick Start Ask the agent to set up or re-register the life dashboard crons and paste the script's output verbatim with its exit status.
