subagent-execution

Delegate eligible steps sequentially to fresh, isolated subagents while the controller maintains central state management, gate control, and final verification. This approach enables scalable execution without sacrificing centralized control over the workflow.

HOTL execution is a governance-backed workflow where independent implementation steps are handed off to isolated workers. You need subagent delegation when a workflow has multiple independent steps and you want to parallelize work while the controller supervises results.

Dispatch a fresh subagent to handle a designated step while the controller retains governance and verification. The controller coordinates runtime execution, manages state under .hotl/state, and enforces non-nested delegation rules.

No, the workflow enforces non-nested delegation. Subagents execute designated steps independently, but critical invariants like verification, gates, and stop conditions remain strictly controller-owned to maintain a single source of truth.

Subagent delegation requires explicit runtime coordination and central state management under .hotl/state. Verification stays in the controller, gates remain controller-owned, and nested delegation is prohibited to ensure governance-backed execution is maintained.

Yes, delegating implementation steps to independent subagents improves throughput while maintaining a single source of truth in the controller. The controller enforces gate control, stop conditions, and final verification over all delegated results.