lev-design（勒夫设计）

lev-design - UX-First Design Pipeline

Lev Concept

What is this? A 7-step agentic workflow that transforms product ideas into validated wireframes. Each step uses specialized frameworks and produces machine-readable artifacts.

Why does it exist? Design is a late-stage compression step. UX teams don't start with screens - they start by understanding problems, jobs, tasks, and information architecture. This skill encodes that process.

When to use it:

• Starting a new feature
• Redesigning existing flows
• Product discovery
• Mobile/desktop app design
• Dashboard/admin UI design

When NOT to use it:

• Simple bug fixes → Just fix it
• Copy changes → Direct edit
• Already have wireframes → Skip to implementation

The 7 Steps

User Request
    ↓
[1. Problem Framing]     ← First Principles, Constraints
    ↓
[2. Jobs to Be Done]     ← JTBD, MBTI User Types
    ↓
[3. Task Decomposition]  ← State Machines, Failure Modes
    ↓
[4. Information Arch]    ← DDD, Entity-Relationship
    ↓
[5. Interaction Models]  ← FSM, Accessibility
    ↓
[6. Component Intent]    ← Atomic Design, I/O Contracts
    ↓
[7. Wireframes]          ← Gestalt, Cognitive Load
    ↓
Design Artifacts + SYNTHESIS.md

CLI Usage

# Auto mode - runs all 7 steps without stopping
/ux "Design the onboarding flow for OpenClaw iOS app"

# Interactive mode - asks questions at each step
/ux full "Design the settings screen"

# Jump to specific step (with prior context)
/ux step 4 --context ./ux-2026-02-01/

# Compare wireframes to current implementation
/ux compare ./ux-2026-02-01/ ./apps/ios/Sources/

Team Structure

Each step spawns:

3 Research Agents (Parallel):

• Codebase patterns (Explore)
• Documentation/decisions (Explore)
• Industry best practices (general-purpose + web)

3 Skill Phases (Sequential):

• Framework Application
• Artifact Generation
• Validation

Artifacts Produced

./ux-{date}/
├─ problem_spec.yaml      # Step 1
├─ jobs.graph.json        # Step 2
├─ task_graph.json        # Step 3
├─ ia_schema.json         # Step 4
├─ interaction_fsm.json   # Step 5
├─ component_intents.yaml # Step 6
├─ wireframe.v1.json      # Step 7
├─ SYNTHESIS.md           # Human summary
└─ COMPARISON.md          # vs current (if exists)

Framework Reference

Integration with Lev Ecosystem

Before design:

• lev get - Gather existing patterns
• lev-research - Industry research

After design:

• design-to-bd - Create epics/tasks
• sidequest - Route implementation by complexity
• lev-builder - POC → Production

During design:

• lev-cdo - Complex architectural decisions
• interview - Framework-powered brainstorming

Expand/Collapse Behavior

Auto-expand when:

• Unresolved questions remain
• Validation found conflicts
• Complexity > 7
• User requested "full" mode

Auto-collapse when:

• High confidence answers
• Clean validation
• Complexity < 4
• User requested "auto" mode

Pivot Detection

Stop and prompt if:

• Fundamental assumption invalidated
• Requirements conflict discovered
• Technical blocker found
• Better alternative revealed by research

Example Usage

> /ux "Design chat interface for OpenClaw iOS"

🔍 Phase 0: Prior Art Check
├─ Found: ChatView.swift, MessageBubble.swift
├─ Found: Previous chat redesign epic (clawd-chat-v2)
└─ Research: ChatGPT, Claude, Gemini UI patterns

📋 Step 1: Problem Framing
├─ Problem: Users need fast, intuitive AI chat on mobile
├─ Constraints: iOS 17+, offline-first, accessibility
└─ Success: <2s response display, 95% task completion

👤 Step 2: JTBD
├─ Job 1: "When I need quick answers, I want to ask naturally"
├─ Job 2: "When context matters, I want to share screenshots"
└─ User types: INTJ (power user), ESFP (casual)

... [Steps 3-7] ...

✅ Complete: ./ux-2026-02-01/
├─ 7 artifacts generated
├─ SYNTHESIS.md ready for review
└─ Comparison: 12 differences from current ChatView

Technique Map

• Role definition - Clarifies operating scope and prevents ambiguous execution.
• Context enrichment - Captures required inputs before actions.
• Output structuring - Standardizes deliverables for consistent reuse.
• Step-by-step workflow - Reduces errors by making execution order explicit.
• Edge-case handling - Documents safe fallbacks when assumptions fail.

Technique Notes

These techniques improve reliability by making intent, inputs, outputs, and fallback paths explicit. Keep this section concise and additive so existing domain guidance remains primary.

Prompt Architect Overlay

Role Definition

You are the prompt-architect-enhanced specialist for lev-design, responsible for deterministic execution of this skill's guidance while preserving existing workflow and constraints.

Input Contract

• Required: clear user intent and relevant context for this skill.
• Preferred: repository/project constraints, existing artifacts, and success criteria.
• If context is missing, ask focused questions before proceeding.

Output Contract

• Provide structured, actionable outputs aligned to this skill's existing format.
• Include assumptions and next steps when appropriate.
• Preserve compatibility with existing sections and related skills.

Edge Cases & Fallbacks

• If prerequisites are missing, provide a minimal safe path and request missing inputs.
• If scope is ambiguous, narrow to the highest-confidence sub-task.
• If a requested action conflicts with existing constraints, explain and offer compliant alternatives.