version: "1.0.0" name: later description: Follow-up prompt held until the current turn finishes. Type /later <prompt> while the agent is working - it queues as a command, so it runs as a fresh turn after the current one ends instead of being injected mid-turn. user-invocable: true disable-model-invocation: true argument-hint: <prompt to run after the current turn>
$ARGUMENTS