cancel

To stop an active automated workflow, you trigger a cancellation command that detects the current state and applies dependency-ordered shutdown logic across all active loops. This immediately halts ongoing work and prevents orphaned processes.

The best way to cancel parallel execution safely is using smart cancellation, which detects active modes and preserves resumable state for specific loops. This allows you to stop team coordination tasks now and continue them later without losing progress.

Yes, force cancellation applies a nuclear cleanup that clears all state completely. This destroys all resumable progress for autopilot and self-improve loops, ensuring a clean slate for all workflow modes.

To resume an interrupted autopilot mode, you rely on the state preservation from a smart cancellation. The system selectively preserves resumable progress for specific modes, allowing you to continue the workflow exactly where you left off later.

Yes, you can parse YAML-defined command arguments like --force or --all to control workflow cancellation behavior. The system uses these arguments to switch between nuclear cleanup and selective state preservation.

You should avoid smart shutdown when you need to completely clear all async pipeline state and start fresh. In this case, smart cancellation preserves state for specific modes, which might conflict with a clean environment requirement.