agentic-engineering-planning

What problem does it solve? Teams often jump from an approved specification straight into coding, producing uncoordinated work, missed dependencies, and unverifiable outcomes. This Skill turns an approved specification, architecture, and design into a single executable plan with explicit task decomposition, ownership, and approval gates before any provisioning or implementation begins. ## Core Features & Use Cases - Task Decomposition and DAG Planning: Breaks an approved spec into a dependency-aware task DAG using the core execution-plan schema, with risks, ownership, and acceptance criteria. - Verification and Rollback Mapping: Maps every planned task to verification steps and rollback procedures so execution can be checked and recovered. - Human Approval Gate: Stops before provisioning or implementation until the named human approval is recorded, preventing premature execution. - Use Case: After your architecture and design docs are approved, invoke this Skill to produce an OpenSpec-backed implementation plan that assigns parallel-safe ownership to agents and defines exactly how each task will be verified. ## Quick Start Use $agentic-engineering-planning to create an OpenSpec-backed implementation plan from the approved specification.