decomposition

Decompose large project tasks into phased, dependency-aware subtasks using topological sorting to enable deterministic parallel multi-agent execution and clear handoffs. This prevents blocked work and reduces coordination overhead by assigning tasks to explicit execution phases.

Dependency resolution uses arrow notation to declare task dependencies and topological sorting to assign execution phases, detecting cycles and prescribing resolution strategies for deterministic multi-agent orchestration. This ensures tasks are sequenced correctly without blocked work.

Structure delegation prompts using spec templates that include objective, context, constraints, acceptance criteria, expected output, and self-improvement steps to ensure clear handoffs. The approach provides examples of strong versus weak prompts and a decision flow for choosing sub-agent versus background-agent delegation.

Task decomposition suits project leads preparing multi-agent plans for feature development, migrations, or cross-team integrations requiring phased execution. It produces topological phase assignments, delegation spec templates, and dependency graphs directly consumable by agents for mixed orchestration patterns.

Cycle resolution strategies are applied during topological phase assignment to detect and break circular dependencies, ensuring deterministic parallel task execution. This prevents blocked work and redundant sequencing by prescribing explicit resolution paths within the dependency graph.