What problem does it solve? Turning an approved PRD, design doc, or experiment proposal into executable work often produces vague task lists with hidden dependencies, missing acceptance criteria, and uncontrolled scope creep. This Skill converts approved intent into an ordered, verifiable task plan that implementers can execute without re-deriving the original intent. ## Core Features & Use Cases - Dependency-aware task graphs: Extract deliverables, map prerequisites and external blockers, and order tasks by dependency and risk so unknowns are de-risked early. - Verifiable task definitions: Every task carries context, acceptance criteria, validation evidence, owner, and stop conditions, plus loop gates defining when work continues, stops, or escalates. - Delivery budget enforcement: Caps implementation at two PRs per issue across repositories, requiring explicit human approval before expanding scope. - Use Case: A developer with an approved feature spec uses this Skill to produce a task plan with execution waves, open blockers, and a handoff package for the shipping agent, complete with telemetry on task counts and plan confidence. ## Quick Start Ask the agent to decompose your approved PRD or design doc into an ordered task plan with acceptance criteria, dependencies, and validation for each task.
