Skip to content

Routing Model

Use this page when you are routing a prompt and want a simple decision model instead of a full catalog scan.

Apple Text uses progressive disclosure: start broad, then move to the narrowest skill that matches the job.

  1. Start with /apple-text:ask.
  2. Decide whether the user needs a workflow, a diagnosis, a choice, or a reference.
  3. Open the narrowest skill that matches that answer shape.

Do not memorize every skill name. Ask whether the prompt is broad, broken, comparative, or already scoped to a subsystem.

  • Broad intake: start with /apple-text:ask. Use this when the problem is clearly Apple text work but still mixed or underspecified.
  • Guided flow or review: start with /skill apple-text-audit. Use this when the user wants an audit, integration walkthrough, or step-by-step implementation path.
  • Broken behavior: start with /skill apple-text-textkit-diag. Use this when the user starts with stale layout, fallback, crashes, or rendering bugs.
  • Tradeoff question: use apple-text-views, apple-text-attributed-string, apple-text-layout-manager-selection. Use this when the main job is picking the right API, architecture, or view.
  • Known API family: start with /skill apple-text-storage or the matching reference skill. Use this when the user already knows the subsystem and wants mechanics, behavior, or API detail.
  • “Why is this editor broken?” -> diagnostic skill
  • “Review this code for risks.” -> workflow skill
  • “Should I use A or B?” -> decision skill
  • “How does this Apple text API behave?” -> reference skill