release: version v3.5.0 - UI Design Workflow with Triple Vision Analysis

🎨 UI Design Workflow System
- Add /workflow:design:auto semi-autonomous orchestrator with interactive checkpoints
- Add /workflow:design:style-extract with triple vision analysis (Claude + Gemini + Codex)
- Add /workflow:design:style-consolidate for token validation and style guide generation
- Add /workflow:design:ui-generate for token-driven HTML/CSS prototype generation
- Add /workflow:design:design-update for design system integration

👁️ Triple Vision Analysis
- Claude Code: Quick initial visual analysis using native Read tool
- Gemini Vision: Deep semantic understanding of design intent
- Codex Vision: Structured pattern recognition with -i parameter
- Consensus synthesis with weighted combination strategy

⏸️ Interactive Checkpoints
- Checkpoint 1: User selects style variants after extraction
- Checkpoint 2: User confirms prototypes before design update
- Pause-and-continue pattern for critical design decisions

🎯 Zero Agent Overhead
- Remove Task(conceptual-planning-agent) wrappers from design commands
- Direct bash execution for gemini-wrapper and codex commands
- Improved performance while preserving all functionality

🎨 Design Features
- OKLCH color format for perceptually uniform design tokens
- W3C Design Tokens compatibility
- WCAG 2.1 AA compliance validation
- Style override support with --style-overrides parameter
- Batch task generation with --batch-plan flag

📚 Documentation Updates
- Update README.md with Phase 2 (UI Design Refinement) section
- Update README_CN.md with Chinese design workflow documentation
- Add comprehensive CHANGELOG entry for v3.5.0
- Add 5 new /workflow:design:* commands to command reference

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
catlog22
2025-10-06 00:06:20 +08:00
parent a18c666f22
commit 25951ac9b0
6 changed files with 456 additions and 152 deletions

View File

@@ -1,45 +1,47 @@
---
name: auto
description: Orchestrate complete UI design refinement workflow from style extraction to brainstorming integration
usage: /workflow:design:auto --session <session_id> --images "<glob>" --pages "<list>" [--interactive] [--variants <count>]
argument-hint: "--session WFS-session-id --images \"refs/*.png\" --pages \"dashboard,auth\" [--interactive] [--variants 2]"
description: Orchestrate UI design refinement workflow with interactive checkpoints for user selection
usage: /workflow:design:auto --session <session_id> --images "<glob>" --pages "<list>" [--variants <count>] [--batch-plan]
argument-hint: "--session WFS-session-id --images \"refs/*.png\" --pages \"dashboard,auth\" [--variants 2] [--batch-plan]"
examples:
- /workflow:design:auto --session WFS-auth --images "design-refs/*.png" --pages "login,register" --interactive
- /workflow:design:auto --session WFS-dashboard --images "refs/*.jpg" --pages "dashboard" --variants 3
allowed-tools: SlashCommand(*), TodoWrite(*), Read(*), Bash(*)
- /workflow:design:auto --session WFS-auth --images "design-refs/*.png" --pages "login,register"
- /workflow:design:auto --session WFS-dashboard --images "refs/*.jpg" --pages "dashboard" --variants 3 --batch-plan
allowed-tools: SlashCommand(*), TodoWrite(*), Read(*), Bash(*), Glob(*)
---
# UI Design Auto Workflow Command
## Overview
Complete autonomous orchestration of UI design refinement workflow: style extraction → consolidation → UI generation → brainstorming integration.
Semi-autonomous UI design workflow with interactive checkpoints: style extraction → **user selection** → consolidation → UI generation → **user confirmation** → design update → optional batch planning.
## Coordinator Role
**Pure orchestrator following /workflow:plan pattern**: Execute 4 design commands in sequence with TodoWrite-driven auto-continuation, no user intervention required.
**Checkpoint-based orchestrator**: Execute Phase 1 automatically, pause for user style selection, execute Phase 3, pause for user prototype confirmation, complete with optional batch task generation.
## Execution Model - Auto-Continue Workflow
## Execution Model - Checkpoint Workflow
This workflow runs **fully autonomously** once triggered:
This workflow runs **semi-autonomously** with user checkpoints:
1. **User triggers**: `/workflow:design:auto --session WFS-xxx --images "refs/*.png" --pages "dashboard,auth"`
2. **Phase 1 executes** (style-extract) → Reports output → Auto-continues
3. **Phase 2 executes** (style-consolidate) → Reports output → Auto-continues
4. **Phase 3 executes** (ui-generate) → Reports output → Auto-continues
5. **Phase 4 executes** (design-update) → Reports final summary
1. **User triggers**: `/workflow:design:auto --session WFS-xxx --images "refs/*.png" --pages "dashboard,auth" [--batch-plan]`
2. **Phase 1 executes** (style-extract) → Reports style cards → **⏸️ PAUSE FOR USER SELECTION**
3. **User selects variants** → Runs `style-consolidate --variants "..."` → Auto-continues
4. **Phase 3 executes** (ui-generate) → Reports prototypes → **⏸️ PAUSE FOR USER CONFIRMATION**
5. **User confirms prototypes** → Runs `design-update --selected-prototypes "..."` → Auto-continues
6. **Phase 5 executes** (batch-plan, optional) → Reports task files
**Auto-Continue Mechanism**:
- TodoWrite tracks current phase status
- After each phase completion, automatically executes next pending phase
- **No user action required** - workflow runs end-to-end autonomously
- Progress updates shown at each phase for visibility
**Checkpoint Mechanism**:
- TodoWrite tracks current phase with "awaiting_user_confirmation" status
- System reports output location and exact command for next step
- User runs provided command to continue workflow
- Workflow resumes automatically after user input
## Core Rules
1. **Start Immediately**: First action is TodoWrite initialization, second action is Phase 1 command execution
2. **No Preliminary Analysis**: Do not read files or validate before Phase 1 (commands handle their own validation)
3. **Parse Every Output**: Extract required data from each command's output for next phase
4. **Auto-Continue via TodoWrite**: Check TodoWrite status to execute next pending phase automatically
5. **Track Progress**: Update TodoWrite after every phase completion
4. **Pause at Checkpoints**: After Phase 1 and Phase 3, pause and prompt user with exact command to continue
5. **User-Driven Continuation**: Workflow resumes when user runs style-consolidate or design-update commands
6. **Track Progress**: Update TodoWrite after every phase completion and checkpoint
## Parameter Requirements
@@ -49,10 +51,10 @@ This workflow runs **fully autonomously** once triggered:
- `--pages "<page_list>"`: Comma-separated list of pages to generate
**Optional Parameters**:
- `--interactive`: Enable interactive style variant selection (default: auto-select first variant)
- `--variants <count>`: Number of UI variants per page (default: 1)
- `--batch-plan`: Auto-generate implementation tasks for selected prototypes after design-update
## 4-Phase Execution
## 5-Phase Execution
### Phase 1: Style Extraction
**Command**: `SlashCommand(command="/workflow:design:style-extract --session {session_id} --images \"{image_glob}\"")`
@@ -60,50 +62,43 @@ This workflow runs **fully autonomously** once triggered:
**Parse Output**:
- Verify: `.design/style-extraction/style-cards.json` created
- Extract: `style_cards_count` from output message
- List available style variant IDs
**Validation**:
- Style cards successfully generated
- At least one style variant available
**TodoWrite**: Mark phase 1 completed, phase 2 in_progress
**TodoWrite**: Mark phase 1 completed, mark checkpoint "awaiting_user_confirmation"
**After Phase 1**: Report to user, auto-continue to Phase 2
**⏸️ CHECKPOINT 1: Pause for User Style Selection**
```
Phase 1 Complete: Style Extraction
Style cards generated: {count}
Available variants: {variant_ids}
Location: .workflow/WFS-{session}/.design/style-extraction/
Continuing to Phase 2: Style Consolidation...
⏸️ USER SELECTION REQUIRED
Review style cards and select your preferred variants.
Then run:
/workflow:design:style-consolidate --session WFS-{session} --variants "{variant_ids}"
Example: /workflow:design:style-consolidate --session WFS-{session} --variants "variant-1,variant-3"
```
---
### Phase 2: Style Consolidation
**Command Construction**:
### Phase 2: Style Consolidation (User-Triggered)
**User Command**: `/workflow:design:style-consolidate --session {session_id} --variants "{selected_variants}"`
```bash
IF --interactive flag present:
command = "/workflow:design:style-consolidate --session {session_id} --interactive"
ELSE:
# Auto-select first variant
command = "/workflow:design:style-consolidate --session {session_id} --variants \"variant-1\""
```
**Command**: `SlashCommand(command="{constructed_command}")`
**Parse Output**:
- Verify: `.design/style-consolidation/design-tokens.json` created
- Extract: `token_count`, `validation_status` from output message
**Validation**:
- Design tokens finalized
- Validation report shows no critical errors
**After user runs command**:
- Workflow automatically continues to Phase 3
- Parse output: token_count, validation_status
**TodoWrite**: Mark phase 2 completed, phase 3 in_progress
**After Phase 2**: Report to user, auto-continue to Phase 3
**Output**:
```
Phase 2 Complete: Style Consolidation
Design tokens: {count}
@@ -115,7 +110,7 @@ Continuing to Phase 3: UI Generation...
---
### Phase 3: UI Generation
### Phase 3: UI Generation (Auto-Triggered after Phase 2)
**Command Construction**:
```bash
@@ -127,41 +122,50 @@ command = "/workflow:design:ui-generate --session {session_id} --pages \"{page_l
**Parse Output**:
- Verify: `.design/prototypes/*.html` files created
- Extract: `prototype_count`, `page_list` from output message
- Extract: `prototype_count`, `page_list`, `generated_files` list
**Validation**:
- All requested pages generated
- HTML and CSS files present for each variant
**TodoWrite**: Mark phase 3 completed, phase 4 in_progress
**TodoWrite**: Mark phase 3 completed, mark checkpoint "awaiting_user_confirmation"
**After Phase 3**: Report to user, auto-continue to Phase 4
**⏸️ CHECKPOINT 2: Pause for User Prototype Confirmation**
```
Phase 3 Complete: UI Generation
Prototypes generated: {count}
Pages: {page_list}
Prototypes generated: {count} ({page_list})
Generated files: {file_list}
Location: .workflow/WFS-{session}/.design/prototypes/
Continuing to Phase 4: Design System Integration...
⏸️ USER CONFIRMATION REQUIRED
Review the generated prototypes and select your preferred variants.
Then run:
/workflow:design:design-update --session WFS-{session} --selected-prototypes "{prototype_ids}"
Example: /workflow:design:design-update --session WFS-{session} --selected-prototypes "dashboard-variant-1,auth-variant-2"
Or to use all generated prototypes:
/workflow:design:design-update --session WFS-{session}
```
---
### Phase 4: Design System Integration
**Command**: `SlashCommand(command="/workflow:design:design-update --session {session_id}")`
### Phase 4: Design System Integration (User-Triggered)
**User Command**: `/workflow:design:design-update --session {session_id} [--selected-prototypes "{selected_prototypes}"]`
**After user runs command**:
- Workflow updates brainstorming artifacts
- If --batch-plan flag present, automatically continues to Phase 5
**Parse Output**:
- Verify: `synthesis-specification.md` updated
- Verify: `ui-designer/style-guide.md` created/updated
**Validation**:
- Brainstorming artifacts successfully updated
- Design system references integrated
**TodoWrite**: Mark phase 4 completed
**Return to User**:
**Output** (if --batch-plan NOT present):
```
UI Design Refinement Complete for session: WFS-{session}
@@ -183,6 +187,45 @@ Next Steps:
(The plan phase will automatically discover and utilize the design system)
```
**Output** (if --batch-plan present):
```
Phase 4 Complete: Design System Integration
Updated Artifacts:
✓ synthesis-specification.md
✓ ui-designer/style-guide.md
Continuing to Phase 5: Batch Task Generation...
```
---
### Phase 5: Batch Task Generation (Optional, Auto-Triggered if --batch-plan)
**Condition**: Only executes if `--batch-plan` flag present
**Execution**:
```bash
FOR each page IN selected_prototypes_pages:
SlashCommand(command="/workflow:plan --agent \"Implement {page} page based on design system\"")
# Parse output task file location
task_files.add(output_location)
```
**TodoWrite**: Mark phase 5 completed
**Return to User**:
```
Phase 5 Complete: Batch Task Generation
Tasks generated for: {page_count} pages
Generated task files:
{task_file_list}
Design workflow complete for session: WFS-{session}
Next: /workflow:execute
```
## TodoWrite Pattern
```javascript

View File

@@ -1,22 +1,23 @@
---
name: style-extract
description: Extract design style from reference images using Gemini Vision and Codex
description: Extract design style from reference images using triple vision analysis (Claude Code + Gemini + Codex)
usage: /workflow:design:style-extract --session <session_id> --images "<glob_pattern>"
argument-hint: "--session WFS-session-id --images \"path/to/*.png\""
examples:
- /workflow:design:style-extract --session WFS-auth --images "design-refs/*.png"
- /workflow:design:style-extract --session WFS-dashboard --images "refs/dashboard-*.jpg"
allowed-tools: Task(conceptual-planning-agent), TodoWrite(*), Read(*), Write(*), Bash(*), Glob(*)
allowed-tools: TodoWrite(*), Read(*), Write(*), Bash(*), Glob(*)
---
# Style Extraction Command
## Overview
Extract design style elements from reference images using a hybrid approach: Gemini Vision for semantic understanding and Codex for structured CSS token generation.
Extract design style elements from reference images using triple vision analysis: Claude Code's native vision, Gemini Vision for semantic understanding, and Codex for structured token generation.
## Core Philosophy
- **Visual Semantic Understanding**: Gemini Vision analyzes design style, typography, and composition
- **Structured Token Generation**: Codex converts semantic analysis to OKLCH + CSS variables + Tailwind config
- **Triple Vision Analysis**: Combine Claude Code, Gemini Vision, and Codex vision capabilities
- **Comprehensive Coverage**: Claude Code for quick analysis, Gemini for deep semantic understanding, Codex for structured output
- **Consensus-Based Extraction**: Synthesize results from all three sources
- **Style Card System**: Generate reusable style cards for consolidation phase
- **Multi-Image Support**: Process multiple reference images and extract common patterns
@@ -31,73 +32,102 @@ EXPAND: glob pattern to concrete image paths
VERIFY: at least one image file exists
```
### Phase 2: Gemini Vision Analysis
**Agent Invocation**: Task(conceptual-planning-agent) with Gemini Vision capability
### Phase 2: Claude Code Vision Analysis (Quick Initial Pass)
**Direct Execution**: Use Read tool with image paths
```bash
Task(conceptual-planning-agent): "
[FLOW_CONTROL]
# Claude Code's native vision capability for quick initial analysis
FOR each image IN expanded_image_paths:
Read({image_path})
# Claude Code analyzes image and extracts basic patterns
Extract visual design semantics from reference images
## Context Loading
ASSIGNED_TASK: style-extraction
OUTPUT_LOCATION: .workflow/WFS-{session}/.design/style-extraction/
IMAGE_INPUTS: {expanded_image_paths}
## Flow Control Steps
1. **load_images_gemini**
- Action: Analyze reference images using Gemini Vision
- Command: bash(cd .workflow/WFS-{session} && ~/.claude/scripts/gemini-wrapper --approval-mode yolo -p \"
PURPOSE: Extract design style semantics from reference images
TASK: Analyze color palettes, typography, spacing, layout principles, component styles
MODE: write
CONTEXT: @{../../{image_paths}}
EXPECTED: JSON with semantic style description (colors with names, font characteristics, spacing scale, design philosophy)
RULES: Focus on extracting semantic meaning, not exact pixel values
\")
- Output: semantic_style_analysis.json
2. **load_session_context**
- Action: Load brainstorming context for style alignment
- Command: Read(.workflow/WFS-{session}/.brainstorming/synthesis-specification.md) || Read(.workflow/WFS-{session}/.brainstorming/ui-designer/analysis.md)
- Output: design_context
- Optional: true
## Analysis Requirements
**Semantic Extraction**: Identify design philosophy (minimalist, brutalist, neumorphic, etc.)
**Color System**: Extract primary, secondary, accent colors with semantic names
**Typography**: Font families, weights, scale (headings, body, small)
**Spacing**: Identify spacing scale patterns
**Components**: Border radius, shadows, button styles, card styles
## Expected Deliverables
1. **semantic_style_analysis.json**: Gemini Vision analysis results
"
# Write preliminary analysis
Write(.workflow/WFS-{session}/.design/style-extraction/claude_vision_analysis.json)
```
### Phase 3: Codex Structured Token Generation
**Agent Invocation**: Task(conceptual-planning-agent) with Codex capabilities
**Output**: `claude_vision_analysis.json` with Claude Code's initial observations
### Phase 3: Gemini Vision Analysis (Deep Semantic Understanding)
**Direct Bash Execution**: No agent wrapper
```bash
Task(conceptual-planning-agent): "
[FLOW_CONTROL]
bash(cd .workflow/WFS-{session}/.design/style-extraction && \
~/.claude/scripts/gemini-wrapper --approval-mode yolo -p "
PURPOSE: Extract deep design semantics from reference images
TASK: Analyze color palettes, typography, spacing, layout principles, component styles, design philosophy
MODE: write
CONTEXT: @{../../{image_paths}}
EXPECTED: JSON with comprehensive semantic style description (colors with names, font characteristics, spacing scale, design philosophy, UI patterns)
RULES: Focus on extracting semantic meaning and design intent, not exact pixel values. Identify design system patterns.
")
Convert semantic style analysis to structured CSS tokens
# Output: gemini_vision_analysis.json
```
## Context Loading
INPUT_ANALYSIS: .workflow/WFS-{session}/.design/style-extraction/semantic_style_analysis.json
OUTPUT_LOCATION: .workflow/WFS-{session}/.design/style-extraction/
TECH_STACK_CONTEXT: CLAUDE.md, .claude/workflows/cli-templates/tech-stacks/*.md
**Output**: `gemini_vision_analysis.json` with Gemini's deep semantic analysis
## Flow Control Steps
1. **load_semantic_analysis**
- Action: Read Gemini Vision analysis
- Command: Read(.workflow/WFS-{session}/.design/style-extraction/semantic_style_analysis.json)
- Output: semantic_analysis
### Phase 4: Codex Vision Analysis (Structured Pattern Recognition)
**Direct Bash Execution**: Codex with -i parameter
2. **generate_css_tokens_codex**
- Action: Convert semantic analysis to structured CSS tokens
```bash
bash(codex -C .workflow/WFS-{session}/.design/style-extraction --full-auto -i {image_paths} exec "
PURPOSE: Analyze reference images for structured design patterns
TASK: Extract color values, typography specs, spacing measurements, component patterns
MODE: auto
CONTEXT: Reference images provided via -i parameter
EXPECTED: Structured JSON with precise design specifications
RULES: Focus on measurable design attributes and component patterns
" --skip-git-repo-check -s danger-full-access)
# Output: codex_vision_analysis.json
```
**Output**: `codex_vision_analysis.json` with Codex's structured analysis
### Phase 5: Synthesis of Triple Vision Analysis
**Direct Execution**: Main Claude synthesizes all three analyses
```bash
# Read all three vision analysis results
Read(.workflow/WFS-{session}/.design/style-extraction/claude_vision_analysis.json)
Read(.workflow/WFS-{session}/.design/style-extraction/gemini_vision_analysis.json)
Read(.workflow/WFS-{session}/.design/style-extraction/codex_vision_analysis.json)
# Load optional session context
Read(.workflow/WFS-{session}/.brainstorming/synthesis-specification.md) [optional]
# Synthesize consensus analysis
# Main Claude identifies common patterns, resolves conflicts, creates unified semantic analysis
Write(.workflow/WFS-{session}/.design/style-extraction/semantic_style_analysis.json)
```
**Synthesis Strategy**:
- **Color system**: Consensus from all three sources, prefer Codex for precise values
- **Typography**: Gemini for semantic understanding, Codex for measurements
- **Spacing**: Cross-validate across all three,identify consistent patterns
- **Design philosophy**: Weighted combination with Gemini having highest weight
- **Conflict resolution**: Majority vote or use context from synthesis-specification.md
**Output**: `semantic_style_analysis.json` - unified analysis synthesizing all three sources
### Phase 6: Structured Token Generation
**Direct Bash Execution**: Codex generates CSS tokens
```bash
bash(codex -C .workflow/WFS-{session}/.design/style-extraction --full-auto exec "
PURPOSE: Convert synthesized semantic analysis to structured CSS design tokens
TASK: Generate W3C-compliant design tokens, Tailwind config, and style card variants
MODE: auto
CONTEXT: @{semantic_style_analysis.json,../../../../CLAUDE.md}
EXPECTED: design-tokens.json (OKLCH format), tailwind-tokens.js, style-cards.json (3 variants)
RULES: $(cat ~/.claude/workflows/design-tokens-schema.md) | OKLCH colors, rem spacing, semantic naming
" --skip-git-repo-check -s danger-full-access)
```
**Output**:
- `design-tokens.json`: W3C-compliant tokens in OKLCH format
- `tailwind-tokens.js`: Tailwind theme extension
- `style-cards.json`: 3 style variant cards for user selection
- Command: bash(codex -C .workflow/WFS-{session}/.design/style-extraction --full-auto exec \"
PURPOSE: Generate structured CSS design tokens from semantic analysis
TASK: Convert semantic color/typography/spacing to OKLCH CSS variables and Tailwind config

View File

@@ -1,12 +1,12 @@
---
name: ui-generate
description: Generate UI prototypes using consolidated design tokens and synthesis specification
usage: /workflow:design:ui-generate --session <session_id> --pages "<page_list>" [--variants <count>]
argument-hint: "--session WFS-session-id --pages \"dashboard,auth,settings\" [--variants 2]"
description: Generate UI prototypes using consolidated design tokens with optional style overrides
usage: /workflow:design:ui-generate --session <session_id> --pages "<page_list>" [--variants <count>] [--style-overrides "<path_or_json>"]
argument-hint: "--session WFS-session-id --pages \"dashboard,auth\" [--variants 2] [--style-overrides \"overrides.json\"]"
examples:
- /workflow:design:ui-generate --session WFS-auth --pages "login,register"
- /workflow:design:ui-generate --session WFS-dashboard --pages "dashboard" --variants 3
allowed-tools: Task(conceptual-planning-agent), TodoWrite(*), Read(*), Write(*), Bash(*)
- /workflow:design:ui-generate --session WFS-dashboard --pages "dashboard" --variants 3 --style-overrides "overrides.json"
allowed-tools: TodoWrite(*), Read(*), Write(*), Bash(*)
---
# UI Generation Command

View File

@@ -5,6 +5,201 @@ All notable changes to Claude Code Workflow (CCW) will be documented in this fil
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
## [3.5.0] - 2025-10-06
### 🎨 UI Design Workflow with Triple Vision Analysis
This release introduces a comprehensive UI design workflow system with triple vision analysis capabilities, interactive user checkpoints, and zero agent overhead for improved performance.
#### Added
**New UI Design Workflow System**:
- **`/workflow:design:auto`**: Semi-autonomous UI design workflow orchestrator
- Interactive checkpoints for user style selection and prototype confirmation
- Optional batch task generation with `--batch-plan` flag
- Pause-and-continue pattern at critical decision points
- Automatic progression between phases after user input
- Support for multiple UI variants per page (`--variants` parameter)
**Triple Vision Analysis Architecture**:
- **Phase 1: Claude Code Vision Analysis**: Quick initial visual analysis using native Read tool
- **Phase 2: Gemini Vision Analysis**: Deep semantic understanding of design intent
- **Phase 3: Codex Vision Analysis**: Structured pattern recognition with `-i` parameter
- **Phase 4: Consensus Synthesis**: Weighted combination by main Claude agent
- **Synthesis Strategy**:
- Color system: Consensus with Codex precision preference
- Typography: Gemini semantic + Codex measurements
- Spacing: Cross-validation across all three sources
- Design philosophy: Weighted with Gemini highest priority
- Conflict resolution: Majority vote or synthesis-specification.md context
**Individual Design Commands**:
- **`/workflow:design:style-extract`**: Extract design styles from reference images
- Triple vision analysis (Claude Code + Gemini + Codex)
- Generates `semantic_style_analysis.json`, `design-tokens.json`, `style-cards.json`
- Outputs multiple style variant cards for user selection
- Direct bash execution (no agent wrappers)
- **`/workflow:design:style-consolidate`**: Consolidate selected style variants
- Validates and merges design tokens
- Generates finalized `design-tokens.json`, `style-guide.md`, `tailwind.config.js`
- WCAG AA compliance validation
- Token coverage ≥90% requirement
- **`/workflow:design:ui-generate`**: Generate production-ready UI prototypes
- Token-driven HTML/CSS generation with Codex
- Support for `--style-overrides` parameter for runtime customization
- Generates `{page}-variant-{n}.html`, `{page}-variant-{n}.css` per page
- Semantic HTML5 with ARIA attributes
- Responsive design with token-based breakpoints
- **`/workflow:design:design-update`**: Integrate design system into brainstorming
- Updates `synthesis-specification.md` with UI/UX guidelines section
- Creates/updates `ui-designer/style-guide.md`
- Makes design tokens available for task generation phase
**Interactive Checkpoint System**:
- **Checkpoint 1 (After style-extract)**: User selects preferred style variants
- Command: `/workflow:design:style-consolidate --session WFS-xxx --variants "variant-1,variant-3"`
- Workflow pauses until user runs consolidation command
- **Checkpoint 2 (After ui-generate)**: User confirms selected prototypes
- Command: `/workflow:design:design-update --session WFS-xxx --selected-prototypes "page-variant-1,page-variant-2"`
- Workflow pauses until user runs design-update command
**Design System Features**:
- **OKLCH Color Format**: Perceptually uniform color space for design tokens
- **W3C Design Tokens Compatibility**: Standard-compliant token format
- **Style Override Mechanism**: Runtime token merging using jq
- **Batch Task Generation**: Automatic `/workflow:plan` invocation for each page
- **Accessibility Validation**: WCAG 2.1 AA compliance checks
#### Changed
**Agent Architecture Simplification**:
- **Removed agent wrappers** from `style-extract` and `ui-generate` commands
- Previously used `Task(conceptual-planning-agent)` for simple bash execution
- Now executes `gemini-wrapper` and `codex` commands directly via Bash tool
- Reduced execution overhead and complexity
- Preserved all functionality while improving performance
**Command Execution Pattern**:
- **Direct Bash Execution**: All CLI tools now use direct bash commands
- Gemini Vision: `bash(gemini-wrapper --approval-mode yolo -p "...")`
- Codex Vision: `bash(codex -i {images} --full-auto exec "..." -s danger-full-access)`
- Codex Token Generation: `bash(codex --full-auto exec "..." -s danger-full-access)`
- No intermediate agent layers
**Workflow Integration**:
- Design phase now optional but recommended for UI-heavy projects
- Seamless integration with existing brainstorming → planning → execution flow
- Design artifacts automatically discovered by `task-generate` if present
- UI tasks automatically include `load_design_tokens` in flow_control
**Updated Documentation**:
- **README.md**: Added UI Design Workflow section in Getting Started
- **README_CN.md**: Chinese documentation updated with design workflow
- **Command Reference**: Added 5 new `/workflow:design:*` commands
- **Phase Renumbering**: Shifted phases to accommodate new Phase 2 (UI Design)
#### Benefits
**User Experience**:
- 🎨 **Visual Validation**: Users confirm design before implementation starts
- ⏸️ **Interactive Control**: Critical design decisions require explicit user approval
- 👁️ **Comprehensive Analysis**: Three AI vision sources provide robust style extraction
- 🎯 **Zero Waiting**: Direct bash execution eliminates agent overhead
- 📦 **Automation Ready**: Optional batch task generation accelerates workflow
**Code Quality**:
- 🔒 **Token Enforcement**: 100% CSS values use custom properties (verified)
-**Accessibility**: WCAG AA validated at design phase
- 🎨 **Consistency**: Single source of truth for visual design (design-tokens.json)
- 🧪 **Production Ready**: Semantic HTML5, responsive, accessible prototypes
**Development Workflow**:
- 🔄 **Seamless Integration**: Optional design phase fits between brainstorming and planning
- 🚀 **Backward Compatible**: Existing workflows unaffected if design phase skipped
- 📊 **Better Planning**: Design system context improves task generation quality
- 🎯 **Focused Implementation**: Developers work from validated prototypes and tokens
#### Technical Details
**Triple Vision Analysis Flow**:
```
Reference Images
Phase 2: Claude Code (Read tool) → claude_vision_analysis.json
Phase 3: Gemini Vision (gemini-wrapper) → gemini_vision_analysis.json
Phase 4: Codex Vision (codex -i) → codex_vision_analysis.json
Phase 5: Main Claude Synthesis → semantic_style_analysis.json
Phase 6: Codex Token Generation → design-tokens.json, style-cards.json
```
**Checkpoint Workflow Pattern**:
```
User: /workflow:design:auto --session WFS-xxx --images "refs/*.png" --pages "dashboard,auth"
Phase 1: style-extract (automatic)
↓ [CHECKPOINT 1: User selects style variants]
User: /workflow:design:style-consolidate --session WFS-xxx --variants "variant-1,variant-3"
Phase 3: ui-generate (automatic after Phase 2)
↓ [CHECKPOINT 2: User confirms prototypes]
User: /workflow:design:design-update --session WFS-xxx --selected-prototypes "dashboard-variant-1,auth-variant-2"
Phase 5: batch-plan (optional, automatic if --batch-plan flag)
```
**Output Structure**:
```
.workflow/WFS-{session}/.design/
├── style-extraction/
│ ├── claude_vision_analysis.json
│ ├── gemini_vision_analysis.json
│ ├── codex_vision_analysis.json
│ ├── semantic_style_analysis.json (synthesis)
│ ├── design-tokens.json (preliminary)
│ └── style-cards.json (variants for selection)
├── style-consolidation/
│ ├── style-philosophy.md
│ ├── design-tokens.json (final, validated)
│ ├── style-guide.md
│ ├── tailwind.config.js
│ └── validation-report.json
└── prototypes/
├── {page}-variant-{n}.html (per page, per variant)
├── {page}-variant-{n}.css (token-driven styles)
├── {page}-variant-{n}-notes.md (implementation notes)
└── design-tokens.css (CSS custom properties)
```
**New Agent Documentation**:
- **`ui-design-agent.md`**: Specialized agent for UI/UX design workflows
- Vision analysis, token generation, prototype creation capabilities
- Multi-modal vision provider strategy (Gemini primary, Codex fallback)
- Quality gates: WCAG AA, token coverage ≥90%, component mapping ≥95%
- Flow control specification for 3 design phases
#### Use Cases
**When to Use Design Workflow**:
- User-facing applications with visual design requirements
- Design system creation and maintenance
- Brand-critical user interfaces
- Projects requiring accessibility compliance
- Multi-page applications with consistent styling
**When to Skip Design Workflow**:
- Backend APIs without UI components
- CLI tools and command-line applications
- Quick prototypes and MVPs
- Projects with existing design systems
---
## [3.4.2] - 2025-10-05
### 📚 CLI Documentation Refactoring

View File

@@ -2,7 +2,7 @@
<div align="center">
[![Version](https://img.shields.io/badge/version-v3.4.2-blue.svg)](https://github.com/catlog22/Claude-Code-Workflow/releases)
[![Version](https://img.shields.io/badge/version-v3.5.0-blue.svg)](https://github.com/catlog22/Claude-Code-Workflow/releases)
[![License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
[![Platform](https://img.shields.io/badge/platform-Windows%20%7C%20Linux%20%7C%20macOS-lightgrey.svg)]()
[![MCP Tools](https://img.shields.io/badge/🔧_MCP_Tools-Experimental-orange.svg)](https://github.com/modelcontextprotocol)
@@ -15,15 +15,16 @@
**Claude Code Workflow (CCW)** is a next-generation multi-agent automation framework that orchestrates complex software development tasks through intelligent workflow management and autonomous execution.
> **🎉 Latest: v3.4.2** - CLI documentation refactoring & single source of truth. See [CHANGELOG.md](CHANGELOG.md) for details.
> **🎉 Latest: v3.5.0** - UI Design Workflow with Triple Vision Analysis. See [CHANGELOG.md](CHANGELOG.md) for details.
>
> **What's New in v3.4.2**:
> - 📚 **CLI Documentation Refactoring**: Eliminated 681 lines of duplicate content across 7 command files
> - 🎯 **Single Source of Truth**: Established implicit reference pattern to `intelligent-tools-strategy.md`
> - 🔍 **Documentation Consistency**: All CLI commands now reference centralized strategy guide
> - 💡 **Maintenance Optimization**: Reduced maintenance overhead while preserving unique command features
> - **Enhanced Clarity**: Streamlined documentation focuses on command-specific capabilities
> - 📖 **Better Organization**: File patterns, templates, and MODE definitions centralized
> **What's New in v3.5.0**:
> - 🎨 **UI Design Workflow**: Complete design refinement workflow from style extraction to prototype generation
> - 👁️ **Triple Vision Analysis**: Combines Claude Code + Gemini + Codex vision capabilities for comprehensive style extraction
> - ⏸️ **Interactive Checkpoints**: User selection points for style variants and prototype confirmation
> - 🎯 **Zero Agent Overhead**: Direct bash execution for CLI tools, removing unnecessary agent wrappers
> - 🎨 **Style Customization**: Runtime override support with `--style-overrides` parameter
> - 📦 **Batch Task Generation**: Optional automatic task creation for selected prototypes with `--batch-plan`
> - 🔄 **Semi-Autonomous Workflow**: User-driven continuation at critical design decision points
---
@@ -250,7 +251,19 @@ MCP (Model Context Protocol) tools provide advanced codebase analysis. **Recomme
/workflow:brainstorm:synthesis # Generate consolidated specification
```
**Phase 2: Action Planning**
**Phase 2: UI Design Refinement** *(Optional for UI-heavy projects)*
```bash
# Extract design styles from reference images and generate prototypes
/workflow:design:auto --session WFS-auth --images "design-refs/*.png" --pages "login,register" --batch-plan
# Or run individual design phases
/workflow:design:style-extract --session WFS-auth --images "refs/*.png"
/workflow:design:style-consolidate --session WFS-auth --variants "variant-1,variant-3"
/workflow:design:ui-generate --session WFS-auth --pages "login,register" --variants 2
/workflow:design:design-update --session WFS-auth --selected-prototypes "login-variant-1,register-variant-2"
```
**Phase 3: Action Planning**
```bash
# Create executable implementation plan
/workflow:plan "Implement JWT-based authentication system"
@@ -259,7 +272,7 @@ MCP (Model Context Protocol) tools provide advanced codebase analysis. **Recomme
/workflow:tdd-plan "Implement authentication with test-first development"
```
**Phase 3: Execution**
**Phase 4: Execution**
```bash
# Execute tasks with AI agents
/workflow:execute
@@ -268,7 +281,7 @@ MCP (Model Context Protocol) tools provide advanced codebase analysis. **Recomme
/workflow:status
```
**Phase 4: Testing & Quality Assurance**
**Phase 5: Testing & Quality Assurance**
```bash
# Generate independent test-fix workflow (v3.2.2+)
/workflow:test-gen WFS-auth # Creates WFS-test-auth session
@@ -341,6 +354,11 @@ MCP (Model Context Protocol) tools provide advanced codebase analysis. **Recomme
|---|---|
| `/workflow:session:*` | Manage development sessions (`start`, `pause`, `resume`, `list`, `switch`, `complete`). |
| `/workflow:brainstorm:*` | Use role-based agents for multi-perspective planning. |
| `/workflow:design:auto` | **NEW** Semi-autonomous UI design workflow with interactive checkpoints for style and prototype selection. |
| `/workflow:design:style-extract` | **NEW** Extract design styles from reference images using triple vision analysis (Claude + Gemini + Codex). |
| `/workflow:design:style-consolidate` | **NEW** Consolidate selected style variants into validated design tokens and style guide. |
| `/workflow:design:ui-generate` | **NEW** Generate token-driven HTML/CSS prototypes with optional style overrides. |
| `/workflow:design:design-update` | **NEW** Integrate finalized design system into brainstorming artifacts. |
| `/workflow:plan` | Create a detailed, executable plan from a description. |
| `/workflow:tdd-plan` | Create TDD workflow (6 phases) with test coverage analysis and Red-Green-Refactor cycles. |
| `/workflow:execute` | Execute the current workflow plan autonomously. |

View File

@@ -2,7 +2,7 @@
<div align="center">
[![Version](https://img.shields.io/badge/version-v3.4.2-blue.svg)](https://github.com/catlog22/Claude-Code-Workflow/releases)
[![Version](https://img.shields.io/badge/version-v3.5.0-blue.svg)](https://github.com/catlog22/Claude-Code-Workflow/releases)
[![License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
[![Platform](https://img.shields.io/badge/platform-Windows%20%7C%20Linux%20%7C%20macOS-lightgrey.svg)]()
[![MCP工具](https://img.shields.io/badge/🔧_MCP工具-实验性-orange.svg)](https://github.com/modelcontextprotocol)
@@ -15,15 +15,16 @@
**Claude Code Workflow (CCW)** 是一个新一代的多智能体自动化开发框架,通过智能工作流管理和自主执行来协调复杂的软件开发任务。
> **🎉 最新版本: v3.4.2** - CLI 文档重构与单一信息源。详见 [CHANGELOG.md](CHANGELOG.md)。
> **🎉 最新版本: v3.5.0** - UI 设计工作流与三重视觉分析。详见 [CHANGELOG.md](CHANGELOG.md)。
>
> **v3.4.2 版本新特性**:
> - 📚 **CLI 文档重构**: 消除 7 个命令文件中的 681 行重复内容
> - 🎯 **单一信息源**: 建立对 `intelligent-tools-strategy.md` 的隐式引用模式
> - 🔍 **文档一致性**: 所有 CLI 命令现在引用集中的策略指南
> - 💡 **维护优化**: 降低维护开销,同时保留独特的命令功能
> - **增强清晰度**: 精简的文档专注于命令特定功能
> - 📖 **更好的组织**: 文件模式、模板和 MODE 定义集中化
> **v3.5.0 版本新特性**:
> - 🎨 **UI 设计工作流**: 从风格提取到原型生成的完整设计精炼工作流
> - 👁️ **三重视觉分析**: 融合 Claude Code + Gemini + Codex 视觉能力进行综合风格提取
> - ⏸️ **交互式检查点**: 用户选择风格变体和原型确认的决策点
> - 🎯 **零智能体开销**: CLI 工具直接 bash 执行,移除不必要的智能体包装
> - 🎨 **风格自定义**: 通过 `--style-overrides` 参数支持运行时样式覆盖
> - 📦 **批量任务生成**: 使用 `--batch-plan` 为选定原型自动创建实现任务
> - 🔄 **半自主工作流**: 在关键设计决策点由用户驱动继续
---
@@ -250,7 +251,19 @@ MCP (模型上下文协议) 工具提供高级代码库分析。**推荐安装**
/workflow:brainstorm:synthesis # 生成综合规范
```
**阶段 2行动规划**
**阶段 2UI 设计精炼** *(UI 密集型项目可选)*
```bash
# 从参考图像中提取设计风格并生成原型
/workflow:design:auto --session WFS-auth --images "design-refs/*.png" --pages "login,register" --batch-plan
# 或者运行单独的设计阶段
/workflow:design:style-extract --session WFS-auth --images "refs/*.png"
/workflow:design:style-consolidate --session WFS-auth --variants "variant-1,variant-3"
/workflow:design:ui-generate --session WFS-auth --pages "login,register" --variants 2
/workflow:design:design-update --session WFS-auth --selected-prototypes "login-variant-1,register-variant-2"
```
**阶段 3行动规划**
```bash
# 创建可执行的实现计划
/workflow:plan "实现基于 JWT 的认证系统"
@@ -259,7 +272,7 @@ MCP (模型上下文协议) 工具提供高级代码库分析。**推荐安装**
/workflow:tdd-plan "使用测试优先开发实现认证"
```
**阶段 3:执行**
**阶段 4:执行**
```bash
# 使用 AI 智能体执行任务
/workflow:execute
@@ -268,7 +281,7 @@ MCP (模型上下文协议) 工具提供高级代码库分析。**推荐安装**
/workflow:status
```
**阶段 4:测试与质量保证**
**阶段 5:测试与质量保证**
```bash
# 生成独立测试修复工作流v3.2.2+
/workflow:test-gen WFS-auth # 创建 WFS-test-auth 会话
@@ -341,6 +354,11 @@ MCP (模型上下文协议) 工具提供高级代码库分析。**推荐安装**
|---|---|
| `/workflow:session:*` | 管理开发会话(`start`, `pause`, `resume`, `list`, `switch`, `complete`)。 |
| `/workflow:brainstorm:*` | 使用基于角色的智能体进行多视角规划。 |
| `/workflow:design:auto` | **新增** 带交互式检查点的半自主 UI 设计工作流,用于风格和原型选择。 |
| `/workflow:design:style-extract` | **新增** 使用三重视觉分析Claude + Gemini + Codex从参考图像提取设计风格。 |
| `/workflow:design:style-consolidate` | **新增** 将选定的风格变体整合为经过验证的设计令牌和风格指南。 |
| `/workflow:design:ui-generate` | **新增** 生成基于令牌的 HTML/CSS 原型,支持可选的风格覆盖。 |
| `/workflow:design:design-update` | **新增** 将最终确定的设计系统集成到头脑风暴产物中。 |
| `/workflow:plan` | 从描述创建详细、可执行的计划。 |
| `/workflow:tdd-plan` | 创建 TDD 工作流6 阶段),包含测试覆盖分析和 Red-Green-Refactor 循环。 |
| `/workflow:execute` | 自主执行当前的工作流计划。 |