windags-decomposer

Decomposing a problem into a validated DAG involves converting a ProblemUnderstanding into a DecompositionResult with nodes, edges, skill assignments, wave definitions, and commitment levels. The process enforces three-pass decisions and ensures topological correctness for execution readiness.

Wave planning in task decomposition organizes DAG nodes into sequential execution waves with assigned commitment levels. You need it when orchestrating end-to-end workflows that require validated topological order and dependency constraints before parallel or sequential execution.

Validating topological order in a task DAG requires enforcing three-pass decision logic and checking dependency constraints during decomposition. This ensures the generated DecompositionResult maintains correct node ordering and valid commitment distributions for wave planning.

DAG decomposition supports vague-node handling through its three-pass decision logic. When converting a ProblemUnderstanding into a DecompositionResult, it processes underspecified nodes and enforces commitment levels to produce a concrete DAG ready for execution.

Assigning skills and commitment levels across DAG nodes is handled during decomposition by converting a ProblemUnderstanding into a structured DecompositionResult. The process validates dependency constraints and distributes commitment levels to ensure topological correctness.

Task decomposition fails dependency validation when the generated DAG violates topological order or commitment distribution constraints. The decomposer enforces three-pass decisions and validates dependencies to ensure the DecompositionResult maintains structural correctness before wave planning.