SuperClaude - Specialized AI Development Framework

Project Overview

SuperClaude is a comprehensive configuration framework specifically designed to enhance the capabilities of Claude Code. It transforms Claude Code from a general AI assistant into a specialized AI development partner, offering 19 specialized commands, 9 cognitive roles, and advanced development methodologies.

Core Features

🎯 Cognitive Role System

SuperClaude provides 9 specialized cognitive roles, each with a specific area of expertise and thought process:

Role	Specialization	Use Case	Example Command
architect	System design, long-term maintainability	Planning large systems	/persona:architect → /user:design --api --ddd
frontend	User experience, visual interface	UI/UX work	/persona:frontend → /user:build --react
backend	Performance, reliability, API	Server-side development	/persona:backend → /user:build --api
analyzer	Root cause analysis, evidence-based decision-making	Debugging complex issues	/persona:analyzer → /user:troubleshoot --investigate
security	Threat modeling, security-first	Security review	/persona:security → /user:scan --security
mentor	Teaching, guided discovery	Learning/Explanation	/persona:mentor → /user:explain --depth
refactorer	Code quality, simplification	Code improvement	/persona:refactorer → /user:improve --quality
performance	Speed optimization, bottleneck analysis	Performance tuning	/persona:performance → /user:improve --performance
qa	Quality assurance, testing	Testing workflow	/persona:qa → /user:test --coverage

🛠️ Specialized Command System

SuperClaude offers 18 specialized slash commands, covering the entire development lifecycle:

Development Commands (3)

• /user:build - General project builder
  • --init New project, --feature Add feature, --react React app, --api Backend, --tdd Test-driven development
• /user:dev-setup - Development environment setup
  • --install Tools, --ci CI/CD, --monitor Monitoring
• /user:test - Testing framework
  • --coverage Coverage, --watch Watch mode, --e2e End-to-end testing

Analysis & Improvement Commands (4)

• /user:analyze - Multi-dimensional analysis
  • --code Code review, --architecture Design, --profile Performance, --security Vulnerabilities
• /user:troubleshoot - Debugging & fixing
  • --investigate Explore, --fix Fix, --five-whys Root cause, --prod Production environment
• /user:improve - Enhance code/performance
  • --quality Refactor, --performance Optimize, --iterate Continuous improvement, --threshold Target percentage
• /user:explain - In-depth explanation
  • --depth Complexity, --visual Diagrams, --examples Code examples

Operations Commands (6)

• /user:deploy - Application deployment
• /user:migrate - Database/code migration
• /user:scan - Security & validation
• /user:estimate - Time/complexity prediction
• /user:cleanup - Project maintenance
• /user:git - Git workflow management

Design & Architecture Commands (1)

• /user:design - System design
  • --api REST, --ddd Domain-Driven Design, --prd Requirements, --openapi Specification, --graphql GraphQL

Workflow Commands (2)

• /user:spawn - Create specialized sub-agents
• /user:document - Document creation

System Commands (1)

• /user:load - Load project context

🔧 MCP (Model Context Protocol) Integration

SuperClaude integrates four powerful MCP servers:

1. Context7 (C7) - Automated documentation lookup

   • Instant lookup of library documentation
   • Seamless integration with GitHub workflows

2. Sequential - Complex analysis

   • Sequential thinking for deep problems
   • Suitable for system design and root cause analysis

3. Magic - UI generation

   • Automatic creation of React components
   • Intelligent UI building

4. Puppeteer - Browser testing

   • Validation of UI components
   • Performance monitoring

🚀 Core Advantages

Token Optimization

• UltraCompressed Mode: 70% Token reduction
• Uses Telegram-style simplification: Removes articles/conjunctions/prepositions
• Symbolization: →(to), &(and), @(at), w/(with), +(add), -(remove)
• Abbreviations: cfg(config), fn(function), impl(implementation)

Research-First Approach

• Automatically looks up official documentation
• Evidence-based decision making
• Citation-driven implementation

Adaptive Thinking

• Adjusts analysis depth based on task
• Intelligent tool selection
• Context-aware responses

Installation and Configuration

Automatic Installation

git clone https://github.com/NomenAK/SuperClaude.git
cd SuperClaude
./install.sh

Manual Installation

# 1. Create structure
mkdir -p ~/.claude/commands/shared

# 2. Copy configuration files
cp CLAUDE.md RULES.md PERSONAS.md MCP.md ~/.claude/

# 3. Copy commands
cp .claude/commands/*.md ~/.claude/commands/

# 4. Copy shared resources
cp .claude/commands/shared/*.yml ~/.claude/commands/shared/

Verify Installation

# Check installation
ls -la ~/.claude/ # Should show 4 main files
ls -la ~/.claude/commands/ # Should show 17 files (16 commands + index)

# Test usage
# Open any project in Claude Code and try:
/user:analyze --code
/persona:architect

Project Structure

~/.claude/
├── CLAUDE.md          # Main configuration (loads other files)
├── RULES.md           # Governance & practices
├── PERSONAS.md        # 9 cognitive archetypes
├── MCP.md             # Model Context Protocol patterns
└── commands/          # 17 specialized slash commands + index
    ├── analyze.md
    ├── build.md
    ├── cleanup.md
    ├── deploy.md
    ├── design.md
    ├── dev-setup.md
    ├── document.md
    ├── estimate.md
    ├── explain.md
    ├── git.md
    ├── improve.md
    ├── index.md
    ├── load.md
    ├── migrate.md
    ├── scan.md
    ├── spawn.md
    ├── test.md
    ├── troubleshoot.md
    └── shared/        # 19 configuration resources
        ├── ambiguity-check.yml
        ├── audit.yml
        ├── checkpoint.yml
        ├── cleanup-patterns.yml
        ├── command-memory.yml
        ├── documentation-dirs.yml
        ├── evidence.yml
        ├── git-operations.yml
        ├── git-workflow.yml
        ├── impl.yml
        ├── loading-cfg.yml
        ├── mcp-flags.yml
        ├── patterns.yml
        ├── performance-monitoring.yml
        ├── planning-mode.yml
        ├── research-first.yml
        ├── thinking-modes.yml
        ├── ultracompressed.yml
        └── validation.yml

Usage Examples

General Flags

General flags supported by all commands:

Planning & Thinking

• --plan - Display execution plan before running
• --think - Standard analysis mode (multi-file context)
• --think-hard - Deep analysis mode (architecture level)
• --ultrathink - Critical analysis mode (system level)

Documentation Control

• --ultracompressed / --uc - ~70% Token reduction mode

MCP Server Control

• --c7 / --no-c7 - Enable/disable Context7 documentation
• --seq / --no-seq - Enable/disable Sequential thinking
• --magic / --no-magic - Enable/disable Magic UI builder
• --pup / --no-pup - Enable/disable Puppeteer browser
• --all-mcp - Enable all MCP servers
• --no-mcp - Disable all MCP servers (native tools only)

Workflow Examples

Full Development Workflow

/persona:architect → /user:load --depth deep → /user:analyze --architecture
/user:design --api --ddd → /user:estimate --detailed
/persona:backend → /user:build --api
/persona:frontend → /user:build --react → /user:test --coverage
/persona:security → /user:scan --security → /user:improve --quality
/user:cleanup --all → /user:deploy --env staging

Debugging Workflow

/persona:analyzer → /user:troubleshoot --investigate --prod → /user:analyze --profile
"think hard about performance bottlenecks" → /user:troubleshoot --five-whys
/persona:performance → /user:improve --performance --iterate --threshold 90%

Testing & Quality Assurance

/persona:qa → /user:test --e2e → /user:scan --validate
/persona:mentor → /user:explain --depth beginner --visual --examples
/user:document --user → /user:build --tdd → /user:analyze --code

Documentation Organization

Operational Docs vs. Project Docs

SuperClaude automatically separates operational reports and project documentation:

.claudedocs/           # Claude operational documentation
├── audit/
├── reports/
├── summaries/
├── metrics/
└── incidents/

/docs                  # Project documentation
├── api/
├── guides/
├── architecture/
└── development/

Automatic Reporting

• analyze → reports/
• scan → reports/
• test → metrics/
• troubleshoot → incidents/

Manual Documentation

• /user:document → API → /docs/api/
• Guides → /docs/guides/

Advanced Features

Model Adaptation

• Simple tasks → sonnet (cost-effective)
• Complex tasks → sonnet-4 (balanced)
• Critical tasks → opus-4 (maximum capability)

Convenience Commands

• /compact - Smart reduction
• /clear - Start fresh
• /model - Switch model

Checkpoint System

A Git-integrated checkpoint system allows developers to navigate back to any point in the workflow:

/user:git --checkpoint before-refactor
/user:git --rollback

Troubleshooting

Common Issues

1. Commands not working: Check ls ~/.claude/commands/ (should show 17 files)
2. Planning issues: Add --plan flag → /user:deploy --plan
3. Role issues: /persona:architect → /persona:security → /user:scan --security
4. MCP issues: "React docs" → C7 | "slow app" → Sequential

Maintenance Commands

• Update: cd SuperClaude && git pull && ./install.sh
• Backup: cp -r ~/.claude/ ~/.claude.backup.$(date +%Y%m%d)
• Uninstall: rm -rf ~/.claude/

Project Benefits

Developer Benefits

• ✅ Consistent workflow across all projects
• ✅ Research-first approach → always consults official documentation
• ✅ Quality standards → built-in best practices
• ✅ Cognitive specialization → the right mindset for every task

Applicable Scenarios

• Developers → Faster, smarter coding workflows
• Teams → Consistent AI-assisted patterns
• Projects → Quality-focused development
• Learning → AI-guided skill development

Technical Specifications

Symbol System

SuperClaude uses a unique symbol system for increased efficiency:

Symbol	Meaning	Abbreviation	Meaning
→	Leads to	cfg	Config
&	And/With	fn	Function
w/	With	impl	Implementation
@	At/Located at	env	Environment
+	Add/Include	dev	Development
-	Remove/Exclude	auth	Authentication
∀	For all/Each	API	Application Programming Interface

Version Information

• Current Version: v4.0.0
• Type: Configuration Framework (non-executable code)
• Compatibility: Claude Code
• Installation Location: ~/.claude/ (global installation)

Summary

SuperClaude transforms Claude Code from a general AI assistant into a specialized development partner, offering:

• A specialized cognitive role system
• A comprehensive command toolkit
• Intelligent MCP integration
• Efficient token optimization
• A research-driven methodology

This framework is particularly suitable for developers and teams who require a consistent, high-quality AI-assisted development experience. Through its modular design and specialized approach, SuperClaude ensures that every development task receives appropriate expertise and tool support.