What problem does it solve? Scheduled jobs on managed hosts often fail silently: symlinked systemd units vanish during git checkouts, unprefixed job names make ownership unclear, and jobs that bypass run recording become invisible to health checks. This Skill enforces conventions that keep every scheduled unit owned, installed safely, and observable. ## Core Features & Use Cases - Directory-and-manifest model: Keeps every schedule in one directory with a manifest listing job name, schedule, command, timeout, and owner alongside the systemd unit files. - Naming and install conventions: Enforces <owner>-<job> prefixes and regular-file installs instead of symlinks, preventing the "Unit to trigger vanished" failure class. - Structured run records: Requires every run to append a JSONL record with timestamps, exit code, duration, and timeout status so downstream health checks can verify execution. - Use Case: When adding a nightly cleanup timer to a Linux host, use this Skill to name it correctly, install it as a regular file with systemctl --user enable --now, and wire run-record emission in the same change. ## Quick Start Use the scheduled-jobs skill to add a new systemd user timer for my nightly backup script with proper naming, installation, and run logging.
