CtrlK
BlogDocsLog inGet started
Tessl Logo

ux-design

UI/UX design patterns: design systems, interaction and motion, onboarding, microcopy, and component-level UX (forms, tables, search, notifications). Topics: component-designer, dashboard-design, design-critique, design-system, empty-states, error-messaging, first-use-onboarding, form-ux-advanced, iconography, micro-interactions, microcopy-writing, mobile-patterns, motion-design, notification-ux, onboarding-ux, search-ux, table-ux, typography-advanced.

SKILL.md
Quality
Evals
Security

ux-design

Router skill. Resolve topic from the request, Read EXACTLY ONE playbook from routes:, then execute it. Never load the whole cluster. [EXPLICIT]

When to use

UI/UX design work: design systems, components, interaction/motion, onboarding, microcopy, and component-level UX (forms, tables, search, notifications). Not for brand/visual identity, frontend code-gen, or product strategy — sibling skills, not here. [INFERENCIA]

In: the request (resolves topic); optional artifact to critique/improve (screen, component, copy, flow); depth. Out: the playbook's deliverable (audit, spec, pattern, or copy), evidence-tagged. [INFERENCIA]

Routing — resolve topic

Map the request to exactly one enum (infer silently; ask only if two fit equally), Read its routes: playbook — one file, nothing else — then: depth=deep → apply exhaustively, verify each step; quick → essentials. Spine: Discover → Analyze → Execute → Validate. [EXPLICIT]

Disambiguation (common collisions): [INFERENCIA]

  • Reusable UI block → component-designer; tokens/scale/foundations → design-system.
  • Error copy → error-messaging; copy elsewhere → microcopy-writing.
  • New-user flow → onboarding-ux; post-clone/cold-start JM-ADK → first-use-onboarding.
  • "No data yet" → empty-states; "review my design" → design-critique.
  • Animation/transition → motion-design; interaction feedback → micro-interactions.
  • Data grid → table-ux; metric layout → dashboard-design; query/filter → search-ux.

Validation gate (acceptance criteria)

Done only when: exactly one playbook was Read; output follows that playbook's shape; every non-obvious claim carries a verification tag (Alfa core set); no {VACIO_CRITICO} unresolved. [EXPLICIT] Quality gates: constitution v6.0.0 (enforcement), evidence tags, script-first rule. [CONFIG] Run-time machinery in assets/ — assets/routing-checklist.md (routing gate) and assets/quality-rubric.json (guardian scoring). [CONFIG]

Anti-patterns

  • Reading 2+ playbooks or skimming the cluster "for context". [EXPLICIT]
  • Answering from general UX knowledge instead of the routed playbook. [INFERENCIA]
  • Guessing topic when ambiguous — ask one crisp question. [INFERENCIA]
  • Green-as-success or invented metrics in critiques. [EXPLICIT]

Self-correct: routed playbook doesn't fit → re-resolve topic once, never force-fit. No enum fits → say so, propose the nearest sibling skill; never fabricate a route. [INFERENCIA]

Repository
JaviMontano/jm-adk-beta
Last updated
First committed

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.