Task Flow Operations

Page scope (task-flow operations niche): This page targets background orchestration operations only (flow lifecycle, cancellation, recovery). For install/upgrade checklists use Upgrading / Migrating. For deployment rollout use Deployment Upgrade Runbook. For release history use Releases.

openclaw flows list
openclaw flows show <flow-id>
openclaw flows cancel <flow-id>

Use these commands to observe flow status and manage long-running orchestration separately from interactive chat turns.

1. List active flows and identify stalled/blocked states.
2. Inspect specific flow details before manual intervention.
3. Cancel only when new scheduling should stop; let active child tasks settle.
4. Re-run status checks and openclaw doctor to validate final state.

• Confirm gateway health and provider reachability first.
• Check whether child tasks are still running before canceling parent flow.
• After cancellation, verify the flow reaches a terminal state cleanly.

For infrastructure failures, continue in Deployment Upgrade Runbook. For config-path breakages, use Upgrading / Migrating.

Cron and webhooks wake the agent on a schedule or external event. Task flow is for multi-step orchestration that may outlive a single chat turn—background children, blocked states, and explicit cancellation. Use flows when you need observability across steps; use cron for simple “run this prompt every morning.”

Official reference: Task flow docs.

Log task IDs when a flow blocks or fails—support threads need the ID to trace children. Cancel stale flows before retrying; duplicate runs can double-charge API usage.