What problem does it solve? Teams often jump into coding before clarifying requirements, architecture decisions, and task dependencies, leading to rewrites, scope creep, and lost institutional knowledge. This Skill enforces a plan-before-code methodology that converts vague ideas into testable constraints, documented decisions, and parallelizable task checklists. ## Core Features & Use Cases - Requirements Engineering: Transforms vague requests like "make it fast" into measurable constraints such as "P99 latency < 200ms" using a structured PRD framework. - Architecture Decision Records: Guides writing ADRs with context, rejected alternatives, and consequences so future engineers understand why choices were made. - Task Decomposition: Breaks epics into dependency-aware task.md checklists with critical paths, enabling frontend and backend teams to work in parallel via interface-first mocks. - Recall-Governed Planning: Loads prior decisions and failures from memory baselines before drafting new plans, preventing repeated mistakes. - Use Case: When starting a new feature like real-time notifications, activate this Skill to produce a PRD with testable constraints, an ADR comparing SSE vs WebSockets vs polling, and a task breakdown identifying the critical path. ## Quick Start Ask the planning-strategy skill to help you plan a new feature by writing a PRD, drafting an ADR for the key technology choice, and decomposing the work into parallelizable tasks.
