spec-planner_skill

Overview

This skill drives dialogue-led specification development using skeptical questioning and iterative refinement. It forces clarification before planning, surfaces hidden constraints, and produces implementation-ready specs with clear trade-offs. Designed to be pragmatic: every response states the current phase and required next actions.

How this skill works

It starts by asking targeted clarifying questions to close gaps in scope, motivation, constraints, edge cases, and success criteria. After clarification, it researches the affected system area and unfamiliar technologies, summarizes findings, and drafts one or more solution options with trade-off analysis and a single recommended approach. The draft is then checked against completeness criteria and refined until all acceptance conditions and risks are addressed.

When to use it

• Planning a new feature or service and you need a concrete implementation plan
• Deciding between architectural options or creating an ADR/RFC
• Scoping work for estimates, sprint planning, or resource allocation
• Evaluating whether a proposed change is worth the effort
• Preparing testable acceptance criteria and ordered deliverables

Best practices

• Always answer at least one round of clarifying questions before requesting a spec
• Be explicit about non-goals to prevent scope creep
• Provide relevant system context or grant read access for discovery work
• Prioritize smallest viable solution, then expand with trade-offs documented
• Treat risks and open questions as deliverables with owners

Example use cases

• 'Spec this out' for a new payment integration, including failure modes and fraud mitigation
• Choose between in-house vs. third-party search with cost/maintenance trade-offs
• Write an RFC for a new eventing system with data contracts and migration plan
• Scope and estimate a cross-service refactor, listing blocked components and effort sizes
• Create an ADR for introducing a new persistence layer and rollback strategy