OpenClaw automations

Automations

This is the automation database: when to use cron or heartbeat, how to split work across sessions or subagents, and how to keep automated lines honest when something stalls.

Automation ruleGood automation is observable, restartable, and explicit about what artifact or handoff should exist next.

Timing and ownership

The core automation choices

Cron vs heartbeatChoose exact-time scheduling vs periodic judgment-based wakes.Sessions, subagents, and handoffsRoute work to the right context and next owner.Reliable automation patternsDesign for receipts, restartability, and anti-false-green behavior.Automation debugging and recoveryRecover the right layer when work stops moving.

Useful examples

Examples of automation jobs worth building

Scheduled site health checkExact-time or regular checks that verify core routes, assets, and response codes.

Artifact watchdogA guard that notices when the next expected file, report, or handoff never appeared.

Queue recovery flowA bounded automation that restarts the right stalled stage instead of reporting drift forever.

Open deeper reference links for this page

Cron vs heartbeat Sessions, subagents, and handoffs Automation debugging and recovery Stalled jobs and runtime drift recovery