All skills
simota avatar

/vision

@c805268
by shingo imotasimota/agent-skills85 stars
15

Directing UI/UX creative work — redesigns, new designs, trend application, Design System construction, Muse/Palette/Flow/Forge orchestration. Use for design direction. Offers a co-design pair mode.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/vision

This session only. Nothing lands on disk.

referencebrand-strategy.md

≈2.2k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Brand Strategy

Reference for Vision's brand recipe. Translate brand identity into UI direction: voice keywords, palette structure, typography pairing, multi-brand orchestration, and brand-fit scoring.


1. Brand Inputs Required

Before drafting any direction, gather:

Input Source Required?
Brand purpose / mission Brand book, founder interviews ✓
Voice & tone keywords (5+) Brand book, marketing copy samples ✓
Anti-keywords (what we are NOT) Repositioning workshop ✓
Primary logo + variants Brand asset library ✓
Existing palette + usage rules Brand book ✓
Existing type system Brand book ✓ if exists
Competitor visual references Compete agent output recommended
Current UI screenshots (3+ pages) Live product ✓ for redesign
Customer personas Field / Cast recommended

If 3+ inputs are missing, request them before proceeding. Do not invent brand identity.


2. Voice Keyword Definition

Define 5 keywords + 5 anti-keywords. Force binary contrasts.

Template

## Brand voice

| Keyword (we are) | Anti-keyword (we are NOT) |
|---|---|
| Direct | Verbose |
| Confident | Boastful |
| Warm | Casual |
| Precise | Rigid |
| Playful | Childish |

Tone applies to: UI copy, marketing, error messages, onboarding.

Each keyword must map to observable UI properties:

Keyword UI manifestation
Direct ≤ 6 words per CTA, no "click here", no exclamation marks
Confident Bold sans-serif, no qualifiers ("might", "perhaps")
Warm Rounded corners (8-12px), soft shadows, friendly emoji in non-critical states
Precise Monospaced numerals, exact units, sentence-case labels
Playful Spring-easing on micro-interactions, optional skeleton-screen Easter eggs

If you can't translate a keyword to UI, it's brand fluff — drop it.


3. Palette Structure

Single-brand product

Primary brand:    1 hue (anchor — logo color)
Primary tints:    9 steps (50-900 scale)
Neutrals:         9 steps (warm gray or cool gray, not pure)
Semantic:         success / warning / danger / info (4 hues × 9 steps)
Surface:          background / surface / surface-elevated
Accent:           1-2 secondary hues for emphasis

Total: ~70-90 raw color tokens, semantic-aliased to ~30-40 tokens.

Brand-fit rules

  • Primary brand color: WCAG 2.2 AA contrast ≥ 4.5:1 against light surface AND dark surface
  • If primary fails contrast, use a tint as the interactive accent (not the hex from the logo)
  • Pure black (#000) and pure white (#FFF) almost never appear; use neutrals (#0F172A / #F8FAFC)

Logo color vs UI color

A logo color is a brand asset; a UI accent is an interaction signal. They overlap, not equate. Document explicitly:

- Logo: #FF5722 — used in logomark, splash, watermark only
- Interactive accent: #E64A19 — same hue, darker, passes 4.5:1 on white

4. Typography Pair

Choose 2 typefaces max for product UI:

Role Typeface Weights Usage
Display Variable sans (e.g., Inter, Geist) 400, 500, 600, 700 Headlines, hero, large numerals
Body Same family or system stack 400, 500 Paragraphs, labels, UI text
(optional) Monospace JetBrains Mono / system mono 400 Code, IDs, exact data

Avoid:

  • Decorative display fonts in product UI (reserve for marketing)
  • 3+ families (incoherent)
  • Weight 100-300 below 16px (illegible at small size)
  • Font superfamilies with mismatched x-heights (visual jitter)

Brand-fit signal: typography choice should evoke 2-3 of the brand voice keywords. "Confident + Direct" → geometric sans (Geist, Inter). "Warm + Playful" → humanist sans (Recoleta, Söhne). "Precise + Editorial" → modern serif + clean sans pair.


5. Multi-Brand Orchestration

For multi-brand products, use Core → Brand → Product cascade:

@core/                         (semantic tokens only — never branded values)
  --color-bg-primary
  --color-text-primary
  --color-action-primary
  --space-md
  --radius-md

@brand-A/                      (brand-specific overrides)
  --color-action-primary: #E64A19   # Brand A orange
  --font-display: 'Inter'

@brand-B/                      (brand-specific overrides)
  --color-action-primary: #2563EB   # Brand B blue
  --font-display: 'Söhne'

@product-X/                    (brand inheritance + product exceptions)
  → inherits from @brand-A
  --color-bg-marketing-hero: linear-gradient(...)  (only on marketing pages)

Rules:

  • Core defines semantic intent (what does this token MEAN), never specific colors
  • Brand layers override values for that semantic token, never add new ones
  • Product layers may add product-specific exceptions but cannot redefine core semantics
  • No flat shared library across brands — that's the "Frankenstein system" anti-pattern

Document the cascade in Guidelines.md.


6. Brand-Fit Scoring

Score existing UI against brand voice on a 5-axis rubric:

Axis Score 1 Score 5
Voice match (copy aligns with keywords) Generic AI copy Distinctive, on-brand voice
Visual identity (logo + palette + type) Mishmash, brand inconsistency Cohesive, recognizable
Tone consistency across surfaces Marketing ≠ product ≠ docs Unified across all touchpoints
Anti-keyword absence Triggers anti-keywords often No anti-keyword violations
Brand differentiation vs competitors Indistinguishable Clearly own visual position

Average ≤ 2.5 → fundamental brand work needed before redesign. 2.5-3.5 → targeted improvements. ≥ 3.5 → polish only.


7. Repositioning Triggers

Before any major redesign, check if brand strategy itself is shifting:

  • Target audience changed (e.g., SMB → enterprise)
  • Market positioning shifted (e.g., budget → premium)
  • Product category expanded (single tool → platform)
  • Competitor landscape shifted (new dominant aesthetic emerged)
  • Customer feedback signals confusion about "who you are"

If any of these triggers fired, brand strategy must precede UI strategy. A redesign on shifting brand foundation produces incoherence.


8. Common Pitfalls

Pitfall Avoidance
"Brand refresh" without new strategy → cosmetic only Force purpose / audience / positioning revisit
Logo color used as interactive accent (fails contrast) Use logo as asset, derive accent from logo hue + WCAG check
3+ typefaces — no hierarchy Cap at 2; lean on weight + size for hierarchy
Voice keywords too abstract ("modern", "innovative") Force concrete, observable keywords
Anti-keywords missing Bias toward what brand IS leaves ambiguity; add 5 NOT-words
Multi-brand: shared flat token library Apply Core → Brand → Product cascade
Brand-fit not measured against competitors Use Compete output; differentiate, don't blend
Marketing brand ≠ product brand One identity, one cascade; allow tone shifts within it
Token names tied to brand color (--color-orange) Use semantic names (--color-action-primary) — survives rebrands
Brand book cited but never grounded in current UI Audit current screens against book; quantify drift

9. Decision Walkthrough Template

Brand inputs gathered:
  □ Purpose / mission
  □ Voice keywords (5)
  □ Anti-keywords (5)
  □ Logo + variants
  □ Existing palette + rules
  □ Type system
  □ Competitor refs (Compete output)
  □ Current UI screenshots (3+)
  □ Personas (Field / Cast)

Voice keywords:
  1. ____  → UI manifestation: ____
  2. ____  → UI manifestation: ____
  3. ____  → UI manifestation: ____
  4. ____  → UI manifestation: ____
  5. ____  → UI manifestation: ____

Anti-keywords:
  1. ____  2. ____  3. ____  4. ____  5. ____

Palette:
  Primary: ____ (passes 4.5:1 on light: ✓ / ✗, dark: ✓ / ✗)
  Interactive accent (if primary fails): ____
  Logo-only color (if different): ____
  Neutrals: warm / cool / pure

Typography pair:
  Display: ____  weights: ____
  Body: ____  weights: ____
  Mono (if needed): ____

Multi-brand cascade needed: ✓ / ✗
  If ✓: brands = ____, products per brand = ____

Brand-fit score (existing UI):
  Voice: ___ / 5
  Visual: ___ / 5
  Tone consistency: ___ / 5
  Anti-keyword absence: ___ / 5
  Differentiation: ___ / 5
  Avg: ___ / 5  → action: cosmetic / targeted / fundamental

Repositioning triggers fired: ____  → block redesign until brand work done?: ✓ / ✗

Handoff:
  → Muse (token system)
  → Prose (voice & tone in copy)

10. References

  • Marty Neumeier, The Brand Gap
  • DTCG (Design Tokens Community Group) v2025.10
  • Brad Frost, Atomic Design — multi-brand cascade
  • Refactoring UI (Adam Wathan & Steve Schoger) — palette generation
  • WCAG 2.2 AA contrast requirements

Source: SKILL.md on GitHub

No alerts14d5 checks · Risk SAFE
  • Gen Agent Trust Hub14d

    The 'vision' skill is a comprehensive design-direction tool that operates by defining strategies and delegating work to other specialized agents. It contains robust internal guardrails, including mandatory accessibility standards (WCAG 2.2 AA) and ethical AI principles. No security risks were identified.

  • Socket14d

    No alerts

  • Snyk14d

    Risk: LOW · No issues

  • Runlayer6mo

    9 files scanned · No issues

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at c805268. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 3 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/vision