What problem does it solve? Repeating a task manually—checking a deploy, rerunning tests, polling a status page—wastes attention and is easy to forget. This Skill lets an agent run any prompt or skill on a fixed schedule or a self-paced dynamic schedule, waking itself via monitored shell output. ## Core Features & Use Cases - Fixed-interval loops: Parse commands like /loop 5m /foo or check deploy every 5m and arm a background shell loop that emits a unique sentinel on each tick. - Dynamic self-pacing: When no interval is given, the agent decides when the next run is worthwhile, arming event-based watchers (git ref changes, log lines, CI completion) plus a fallback heartbeat timer. - Lifecycle management: Tracks loop PIDs, avoids duplicate loops, runs the prompt once immediately, and cleanly stops loops and watchers on request. - Use Case: Tell the agent /loop 10m run tests during a debugging session; it reruns your test suite every ten minutes and reports what changed after each tick until you ask it to stop. ## Quick Start Ask the agent to run a specific prompt every five minutes using the loop command, for example by typing /loop 5m check deploy status.
