implement

Wave-based plan execution works by dispatching parallel agents across multiple waves to implement assigned files independently. The orchestrator coordinates the builds and verification while maintaining agent isolation to ensure clean integration and fresh context per agent.

The best way to orchestrate parallel agents is using a controlled governance model where an orchestrator handles coordination and dispatches waves. Agents perform the actual code generation, build, and verification tasks, then report back to maintain lean orchestration.

You execute a feature plan by triggering the wave-based execution flow, which loads the plan specification and dispatches parallel agents. Each wave builds and verifies independently, applying structured plan execution across multiple tasks while maintaining isolation.

You need wave-based parallel agents when implementing large feature plans that require fresh context per agent and clean integration. This approach ensures each wave builds and verifies independently, satisfying requirements for plan loading and final verification.

Yes, plan-driven execution maintains agent isolation by dispatching parallel agents across waves with fresh context. Each wave builds and verifies independently, ensuring clean integration while the orchestrator handles coordination and final verification governance.