← ClaudeAtlas

loom-prototype-and-spikelisted

Use when the next step is building a disposable prototype/spike to answer a specific design, state-model, UI, interface, or integration question before committing production implementation.
z3z1ma/agent-loom · ★ 15 · Web & Frontend · score 80
Install: claude install-skill z3z1ma/agent-loom
# loom-prototype-and-spike Prototype and spike is an exploration playbook. It builds the smallest disposable artifact that answers a question, records the answer, and then deletes, absorbs, or routes the result. ## Loom Routing Common routes use these Loom skills for durable records or follow-up workflow: `loom-research`, `loom-evidence`, `loom-specs`, `loom-constitution`, `loom-knowledge`, `loom-tickets`, and `loom-audit`. Follow any named Loom skill fully. This playbook adds workflow pressure; it does not shorten target-skill requirements. Prototype code is not production work unless a later ticket deliberately absorbs it. A prototype is not a substitute for shaping the direction. Use it only after the question is specific enough to answer without deciding the whole product, system shape, data model, state model, or design direction by accident. ## Use This Playbook When Use this playbook when: - the operator asks to prototype, spike, sanity-check, or try options - a state machine, data model, or information model is hard to reason about on paper - UI direction needs visible alternatives before implementation - interface shape is uncertain and examples would clarify it - an integration assumption needs fast proof before planning - a design decision would be expensive to reverse after full implementation Skip it when existing source, specs, or tests can answer the question cheaply. ## Route Use this route: ```text question -> branch -> build throwaway -> observ