GitHub Actions schedule disabled after 60 days: a Sume Scheduled fix
A public repo's scheduled workflows are disabled after 60 days without activity. How Sume Scheduled's active and inactive status differs.

GitHub disables scheduled workflows in a public repository after 60 days with no repository activity, so a quiet automation can stop silently. Sume Scheduled has no inactivity timer: a schedule is active or inactive, and only a person changes that. Whichever you use, add a check that proves the last run happened.
What does the GitHub page say?
The events reference lists the rule next to the other schedule limits.
| Topic | What the page says |
|---|---|
| Inactivity | In a public repository, scheduled workflows are disabled after 60 days without repository activity |
| Branch | Runs on the default branch only |
| Shortest interval | 5 minutes |
| Delays | Can be delayed under high load |
How does Sume Scheduled handle on and off?
A schedule has a status of active or inactive. Inactive means off, and the dashboard toggle is the only thing that turns it off or on. An API start against an inactive schedule is refused with 409 action_inactive, so a paused schedule never runs by accident. The trigger type, cron or api, is fixed when you create it. A cron schedule can also enable the API trigger.
I could not find an automatic disable rule in the Sume docs, and I will not guess at one. The honest statement is narrower: the docs describe status as a setting, not as a timer.
How do I notice a silent stop either way?
Treat a missing run as the alarm, not a failed one. Each run leaves a receipt you can read with GET /v1/action-runs/{id}, and its trigger source says whether it came from cron, manual or api. A webhook for action.run.terminal tells you a run ended, and a run that never starts sends nothing, so the check has to live outside the run.
- Keep a daily check that lists the latest receipt time and alerts if it is older than your interval plus a margin.
- If you drive Sume from a GitHub workflow, make a trivial commit or use a separate keep-alive path you trust, and still alert on the missing receipt.
- Remember that a skipped run (previous run active) is a normal 200 with status skipped, not a failure.
Where to read next
The scheduled agents guide covers status and the spend cap, and the API trigger reference lists the 409 codes. Run webhooks are covered in the webhook guide.
Sources
Related posts
More in Agents
- Higgsfield MCP has no credit cap: cap spend with Sume max_spend_usd
Higgsfield's MCP guide says there is no built-in credit spending cap. Sume's MCP tools take dry_run and max_spend_usd. How they work, and what they skip.
- hypit Understand order: one probe, then parallel batches
Sume's hypit Understand order: probe alone, then transcribe, boundaries and tiles in one batch, then notes. Which verbs wait on the transcript.
- How an agent picks a video model over hosted MCP
An agent reads video-router_models, picks an id, then calls generate_video with an idempotency_key and waits with jobs_wait. Omit the model for sume/auto.
- Get a typed podcast clip plan from an Agent Completion
Send a transcript and an output_schema to POST /v1/agent/completions, and get back start and end times for each clip, ready for video-trim.
Written by Sume