Skill

ux-design-process

UX design intelligence for code review and planning. Provides heuristic evaluation checklists (Nielsen Norman 10 + Baymard 207 adapted), interaction pattern libraries, flow validation, cognitive walkthrough protocols, and aesthetic quality scoring. Routes to greenfield (new project) or brownfield (existing codebase) UX processes based on context. Trigger keywords: UX, usability, heuristic evaluation, user experience, interaction design, cognitive walkthrough, accessibility, touch targets, user flow, design review, UX audit, Nielsen, Baymard, usability testing.

From rune
Install
1
Run in your terminal
$
npx claudepluginhub vinhnxv/rune --plugin rune
Tool Access

This skill uses the workspace's default tool permissions.

Supporting Assets
View in Repository
CREATION-LOG.md
references/aesthetic-direction.md
references/brownfield-process.md
references/greenfield-process.md
references/heuristic-checklist.md
references/interaction-patterns.md
references/ux-pattern-library.md
references/ux-scoring.md
references/web-interface-rules.md
Skill Content

/rune:ux-design-process -- UX Design Intelligence

UX-informed code review and planning support. Provides structured heuristic evaluation, interaction pattern validation, flow analysis, and aesthetic quality assessment.

Load skills: frontend-design-patterns, context-weaving

Configuration

UX design intelligence is gated by talisman configuration:

# .rune/talisman.yml
ux:
  enabled: false            # Default: false (opt-in for initial release)
  cognitive_walkthrough: false  # Default: false (expensive -- uses Opus model)
  touch_target_minimum: 44  # px -- 48px recommended
  finding_prefix: true      # Emit UXH/UXF/UXI/UXC prefixed findings
  blocking: false           # UX findings are non-blocking by default

Important: ux.enabled defaults to false. Users must explicitly opt in via talisman.yml. When the ux: section is missing entirely from talisman.yml, all UX features default to disabled (enabled: false, cognitive_walkthrough: false, blocking: false, touch_target_minimum: 44, finding_prefix: true).

Process Routing

Determine project context and route to the appropriate UX process:

1. Check talisman.yml: ux.enabled === true
   - If false: exit with "UX intelligence is disabled. Enable via talisman.yml: ux.enabled: true"

2. Determine project context:
   - --greenfield flag OR no existing components/ directory → greenfield process
   - --brownfield flag OR existing components/ directory   → brownfield process
   - --audit flag                                          → heuristic audit (both processes)

3. Route:
   - Greenfield → see [greenfield-process.md](references/greenfield-process.md)
   - Brownfield → see [brownfield-process.md](references/brownfield-process.md)
   - Audit      → see [heuristic-checklist.md](references/heuristic-checklist.md)

UX Finding Prefixes

All UX findings use standardized prefixes for integration with the Rune dedup hierarchy:

PrefixDomainAgentDescription
UXHHeuristicux-heuristic-reviewerNielsen/Baymard heuristic violations
UXFFlowux-flow-validatorUser flow breakages, dead ends, loops
UXIInteractionux-interaction-auditorMicro-interaction and state issues
UXCCognitiveux-cognitive-walkerCognitive load and learnability issues

Severity levels per finding:

LevelMeaningExample
P0Blocks user flowForm submit does nothing, nav leads to 404
P1Degrades experienceMissing loading state, no error recovery
P2Cosmetic / improvementInconsistent spacing, suboptimal label text

See ux-scoring.md for the full scoring rubric.

Architecture (3 Layers)

Layer 1: UX Research & Context  -- Understand user needs and project constraints
Layer 2: Evaluation & Analysis  -- Apply heuristics, validate flows, audit interactions
Layer 3: Scoring & Integration  -- Score findings, integrate with Rune review pipeline

Layer 1: UX Research & Context

Greenfield Projects

For new projects, the UX process covers user research, persona development, information architecture, wireframing guidance, and visual design principles.

See greenfield-process.md for the full greenfield workflow.

Brownfield Projects

For existing codebases, the UX process covers heuristic evaluation, usability audit, incremental improvement workflow, design debt assessment, and progressive enhancement.

See brownfield-process.md for the full brownfield workflow.

Layer 2: Evaluation & Analysis

Heuristic Evaluation

50+ checklist items organized by Nielsen Norman's 10 heuristics, adapted from Baymard's 207 usability guidelines for code-level review context. Each item includes an ID, category, description, code-level check instruction, and severity weight.

See heuristic-checklist.md for the full checklist.

UX Pattern Library

Catalog of UX patterns for agent guidance: loading states (skeleton, spinner, progressive), error handling (inline, toast, modal, boundary), form validation, navigation patterns, empty states, confirmation dialogs, and undo patterns.

See ux-pattern-library.md for the pattern catalog.

Interaction Patterns

Micro-interaction and state transition patterns: hover/focus/active states, loading transitions, error recovery flows, optimistic updates, debounce/throttle, gesture handling, scroll behavior, and drag-and-drop feedback.

See interaction-patterns.md for interaction patterns.

Aesthetic Direction

Visual design principles for code-level enforcement: color coherence, typography hierarchy, whitespace rhythm, visual weight balance, and contrast ratios (WCAG AA compliance).

See aesthetic-direction.md for aesthetic principles.

Web Interface Rules

Adapted from Vercel Web Interface Guidelines: semantic HTML, keyboard accessibility, touch target sizes, responsive breakpoints, animation performance, reduced motion preferences, and color-blind safe palettes.

See web-interface-rules.md for interface rules.

Layer 3: Scoring & Integration

Finding Severity

UX findings follow the UXH/UXF/UXI/UXC prefix scheme and are scored P0-P2. UX prefixes are positioned below FRONT in the Rune dedup hierarchy.

See ux-scoring.md for the scoring framework.

Integration with Rune Review Pipeline

UX review agents participate in the Roundtable Circle as conditional Ashes:

Activation: talisman.yml → ux.enabled === true
Trigger: changed files include frontend components (*.tsx, *.jsx, *.vue, *.svelte)
Position: After standard Wave 1 Ashes, before Wave 2 investigation
Blocking: false (by default -- configurable via talisman.yml → ux.blocking)

Cognitive Walkthrough (Optional)

When ux.cognitive_walkthrough: true, the ux-cognitive-walker agent performs a step-by-step cognitive walkthrough of user flows in changed components. This is expensive (uses Opus model) and is OFF by default.

See heuristic-checklist.md for cognitive walkthrough items.

Cross-References

References

Stats
Parent Repo Stars1
Parent Repo Forks0
Last CommitMar 18, 2026