UX Brand

UX Brand

Workflow

1. Parse $ARGUMENTS → {subcommand} + {subject} + options. Default subcommand = brand. Missing subject → AskUserQuestion.

2. Discover inputs — Glob in priority order:

   • biz/{subject}/ux/*-{subject}.md (empathy-map, persona, forces, jtbd, vpc, journey-map)
   • biz/PRD-*-llm.md, biz/analysis/*{subject}*
   • biz/{subject}/ux/design-system.yaml (existing tokens — required for audit and --update)
   • Nothing found → AskUserQuestion: audience description, industry, emotional need, tech-comfort (1-10)

3. Classify audience — See references/brand-archetypes.md:

   • Extract emotional landscape from empathy map feels + forces anxiety/push → select primary archetype (+ optional blend)
   • Extract tech-comfort from persona tech_comfort field (or infer from tools_used, age, profession)
   • --tone and --tech-comfort override artifact-derived values
   • Detect locale/industry from artifact content + --industry hint

4. Research references — If WebSearch available:

   • Generate queries from archetype × tech-comfort × industry. See references/benchmark-framework.md
   • Discover 4-layer references: familiarity anchors, aspiration ceiling, anti-references, cross-industry
   • Fallback: read scripts/benchmark-defaults.yaml

5. Derive tokens — See references/brand-archetypes.md + references/color-psychology.md + references/typography-strategy.md:

   • Select palette/typography/iconography base from archetype
   • Adjust for tech-comfort level (see matrix)
   • Adjust for cultural context (references/cultural-context.md)
   • Test all color pairs for WCAG AA (references/color-psychology.md §WCAG)
   • Validate against references/anti-patterns.md
   • Derive visual/imagery strategy — When tech-comfort ≤ 5 OR early-stage product (no asset budget), default imagery.strategy to zero-asset and derive:
     • Emoji vocabulary: map each UI concept (statuses, actions, alerts, categories) to one consistent emoji. Semantic only, never decorative.
     • Color blocks: hero gradient (dark → primary), alternating section backgrounds, stat cards as visual anchors (large numbers in 48px+ bold = the "imagery")
     • Typography as design: giant stat numbers replace illustrations for feature explanation
     • Whitespace as design: generous spacing replaces decorative elements
     • Set imagery.upgrade-path to guide future investment (e.g., "v2: commission line illustrations")
   • Output schema: references/design-system-schema.md

6. Generate outputs — Branch by subcommand:

brand (default) — Full identity derivation

Write 3 artifacts:

• biz/{subject}/ux/design-system.yaml — tokens (superset of /ux-wireframe format, backward compatible). If --update: read existing tokens, compute diff, update incrementally (not replace). Include "Design System Diff" section in rationale doc.
• biz/{subject}/ux/brand-sheet-{subject}.html — self-contained style tile: palette with hex + contrast ratios, typography scale, component samples (button/card/input/alert/badge), do/don't visual comparisons, reference mood, archetype badge. Inline CSS, no CDN. See references/rationale-template.md §Brand Sheet for content spec. When imagery.strategy = "zero-asset", include a visual strategy demo section at the top: hero gradient mock (live CSS), stat card pattern (3 cards with large numbers), and emoji semantic mapping table. This shows downstream skills and stakeholders HOW the zero-asset approach looks in practice, not just what the tokens are.
• biz/{subject}/ux/brand-rationale-{subject}.md — decision log per references/rationale-template.md

audit — Evaluate alignment

Read existing biz/{subject}/ux/design-system.yaml + UX artifacts. Report:

• Archetype alignment score (derived vs actual tokens)
• Tech-comfort mismatches (e.g., icon-only for low-tech)
• Contrast failures, palette coherence, font appropriateness
• Recommendations with rationale

references — Research only

Present 4-layer reference analysis to console (no files written):

• Familiarity anchors + aspiration ceiling + anti-references + cross-industry
• Each with URL, rationale, and relevance to archetype × tech-comfort

7. Open — xdg-open / open the brand sheet (brand subcommand) or print report (audit/references).

8. Human gate — Present summary to user: archetype selected, tech-comfort level, key palette colors, font choice, reference sites. AskUserQuestion: "Does this brand identity direction look right? Adjust anything before finalizing?" Wait for approval. Downstream skills (/ux-wireframe, /ux-polish) consume these tokens — changes after are expensive.

Archetype → Design Quick Ref

Testing

• Standalone: Smoke test with /ux-brand relance-impayes against existing UX artifacts, then verify downstream chain /ux-wireframe → /ux-polish reads generated tokens.