product-manager

Feature decomposition organizes large features into self-contained phases with explicit inputs, outputs, and dependencies. This Skill applies a DAG structure to generate a YAML manifest that defines phase boundaries, acceptance criteria, and testable work items, enabling parallel development where dependencies allow.

A directed acyclic graph (DAG) represents phase relationships and dependencies without circular loops. DAG-structured planning identifies which phases can run in parallel and which must sequence, reducing bottlenecks and clarifying critical path dependencies in feature delivery.

Provide a feature prompt describing what you need built. The Skill generates a DAG of phases with explicit inputs, outputs, and acceptance criteria, then outputs a YAML manifest detailing phase boundaries, hard and soft dependencies, and concrete deliverables for each phase.

The Skill produces a YAML manifest that structures phases with inputs, outputs, acceptance criteria, and traceable work items. This format captures phase-level deliverables and dependency relationships, making the plan executable and testable by development teams.

Phases are self-contained and testable. The DAG defines both hard dependencies (strict sequencing required) and soft dependencies (preferred but not blocking), so teams can identify which phases run in parallel and which must wait, optimizing delivery timeline.