task-management

Multi-agent task coordination in ChaHua rooms is managed via structured orchestration workflows that decompose goals and assign work to prevent stalling. The system tracks progress, verifies completion against original requirements, and runs retrospectives to adjust plans upon failure.

Goal decomposition breaks complex objectives into executable, assignable steps with clear acceptance criteria and dependencies. This structured breakdown allows work to be distributed to appropriate room agents based on their names, recent performance, and capabilities.

Concurrent background task execution for independent workstreams is supported in ChaHua rooms. The orchestration workflow manages these parallel agent runs while persisting task artifacts to the ./task directory to maintain state across concurrent operations.

Progress tracking and completion verification monitors step status and evaluates final output against original goal requirements. When goals are not met, it runs failure retrospectives to adjust plans, requiring user confirmation for task state changes via propose interfaces.

The task management workflow activates only when a <current_task> context block is present in the ChaHua room. This context block triggers the structured orchestration workflow to manage active room tasks requiring goal decomposition, role-aligned work assignment, or progress tracking.

User confirmation is required to change task states during execution via propose interfaces. This ensures users maintain control over stepwise task management transitions, artifact persistence to the ./task directory, and failure-driven plan adjustments.