OpenClaw Cron Delivery Failure Fix: Restore Scheduled Reports

Teams often discover a cron delivery failure only after a report is missing, even though the automation may have executed somewhere in the background. This search usually appears after the first OpenClaw demo feels promising but the rollout still feels risky. The question is no longer whether an agent can answer a message. The question is whether it can run a real operating lane with memory, permissions, routing, verification, and a clean handoff back to people.

30-second answer

Separate execution from delivery. Use openclaw cron show to inspect the resolved route, openclaw cron runs to inspect history, and channel health checks to confirm the target can receive messages. A job that ran is not the same as a job that reported where humans expected.

When this is worth doing

This is a high-intent troubleshooting page because scheduled agents become invisible infrastructure. If delivery is wrong, the business loses the safety net that made the cron valuable in the first place.

Official docs to keep open

This guide stays inside the documented OpenClaw surface. The most relevant docs are cli/cron.md; automation/cron-jobs.md; channels/channel-routing.md; channels/troubleshooting.md; gateway/health.md. The building blocks to evaluate are cron delivery preview; failure delivery; skipped run behavior; channel routing; Gateway health. If a workflow would need a hidden feature, a private API, or an assumed limit that the docs do not describe, keep it out of the first rollout.

Buyer-intent runbook

1. Run openclaw cron show for the job and read the resolved delivery route. The CLI docs explain that delivery can resolve through channel state and ownership rules.
2. Inspect openclaw cron runs for execution status. A skipped run, failed run, and successful run with bad delivery are different fixes.
3. Check the target channel with openclaw channels status --probe or the channel-specific troubleshooting ladder. Channel auth can fail after the cron definition stays unchanged.
4. Confirm failure alerts are configured where the operator will actually see them. Failure delivery that lands in the wrong place is just another silent failure.
5. Make one narrow change and force or wait for one safe test run. Do not rewrite the prompt, model, schedule, and delivery target at the same time.

Proof before rollout

The proof is a cron show output with the expected route, run history showing execution, and a message in the intended channel or failure alert path. If any one of those is missing, the fix is incomplete.

Common mistakes

• Do not call a cron healthy because it exists in the list.
• Do not ignore skipped runs when the job has a time-sensitive purpose.
• Do not rely on last-channel delivery for important operations unless the route is verified.
• Do not send duplicate manual summaries when one verified completion update is enough.

Rollout note

For every important cron, document owner, schedule, delivery target, failure alert target, and the verification command. That tiny contract saves hours later.

Where the Playbook helps

The Playbook includes the operational habit most teams miss: scheduled work is not complete until delivery and failure reporting are verified. The OpenClaw Playbook turns that decision into a repeatable operating system: which files to keep, which jobs to schedule, which approvals to require, and how to report proof without flooding the team. If you are moving from experiment to revenue or client operations, use the Playbook before the agent becomes another unmanaged tool.

The practical rule is to start with one lane, one owner, one channel, and one verification habit. Delivery is part of the product. A cron that computes the right answer but hides it from the owner is not a working cron. That keeps the first deployment measurable. It also gives the team a simple before-and-after comparison: how long the workflow took manually, what the agent handled, what still needed judgment, and which check proved the result. Once the lane is stable, duplicate the pattern for adjacent work instead of designing a giant automation program on day one.