scope-new-skilllisted

# Scope a new skill Use this skill to explore a possible skill before creating or editing skill files. The goal is to turn a vague intent into a clear implementation shape, or to decide that a skill is not warranted. Mental model: **screen the intent**. Run the candidate through a fixed panel of readiness dimensions (Phase 4) and report a verdict; only intents that survive the screen are admitted as skills, and a failed screen is a valid, informative result. Governing intuition: **decline is a first-class outcome** — explore to a clear implementation shape or to a well-founded "do not create a skill"; never rubber-stamp a skill into existence. **Input**: the proposed skill idea or workflow to evaluate — if the intent is too vague to classify safely, ask one clarifying question (Phase 1) before proceeding. **Boundary**: explores and recommends a shape; does not create or edit skill files unless the user asks to implement after the exploration. ## Workflow ### Phase 1: Intent Capture Identify the user's proposed capability in one sentence. Extract: - The task or situation that would trigger the skill. - The expected user-visible output or action. - The agent hosts or environments the skill should support. - Any known files, tools, APIs, policies, or project contexts involved. Ask a clarifying question only when the intent cannot be safely classified from context. ### Phase 2: Type Classification Read [references/skill-types.md](references/skill-types.md) when the typ