Problem

When the gateway restarts after a cron job's scheduled time has passed, the job is skipped and waits for the next cycle. This causes missed runs if the gateway happens to be down during a job's scheduled time.

Example scenario:

• Job scheduled for 08:00 daily
• Gateway restarts at 09:35 (e.g., after an update or system wake)
• The 08:00 job is skipped, nextRunAtMs jumps to tomorrow
• User misses their daily digest/report

Proposed Solution

Add an optional catchUp flag to cron schedules. When enabled:

1. On gateway startup, compare lastRunAtMs with the expected run time
2. If the job was supposed to run since the last execution but didn't, trigger it immediately
3. Then resume normal scheduling

Suggested Config

{
  "schedule": {
    "kind": "cron",
    "expr": "0 8 * * *",
    "tz": "Asia/Shanghai",
    "catchUp": true
  }
}

Behavior

Alternatives Considered

• Heartbeat-based workaround: Check cron status in HEARTBEAT.md and manually trigger missed jobs. Works but adds complexity and token cost.
• Always catch up: Could be surprising for some use cases, hence making it opt-in.

Additional Context

This is especially important for daily digest/report jobs where missing a day means missing that day's content entirely.