8000
Skip to content

stocktonellis/claude-config

Β 
Β 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

8 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

Claude Agent System

A comprehensive collection of specialized Claude agents and orchestration tools for complex development workflows. This system enables automated multi-agent coordination for everything from feature development to bug investigation.

πŸš€ Quick Start

For Complex Tasks (Recommended)

Use the Agent Orchestrator to coordinate multiple agents automatically:

@agent-orchestrator: Implement user authentication with Firebase. Include email/password login, OAuth with Google, session management, proper error handling, TypeScript compliance, and comprehensive tests.

For Simple Tasks

Call specific agents directly:

@agent-developer: Fix the TypeScript error on line 45 of login.tsx
@agent-quality-reviewer: Review the authentication service for security issues
@agent-technical-writer: Document the new API endpoints

πŸ“ Project Structure

agents/                    # Specialized agents
β”œβ”€β”€ agent-orchestrator.md  # Multi-agent workflow coordinator
β”œβ”€β”€ architect.md           # Solution design and architecture
β”œβ”€β”€ codebase-finder.md     # Code pattern location
β”œβ”€β”€ debugger.md            # Systematic bug investigation
β”œβ”€β”€ developer.md           # Code implementation
β”œβ”€β”€ feature-enhancer.md    # Enhance existing features
β”œβ”€β”€ quality-reviewer.md    # Quality validation and review
└── technical-writer.md    # Documentation creation

commands/                  # Slash commands
└── orchestrate.md         # Universal task orchestration command

workflows/                 # Detailed workflow templates
β”œβ”€β”€ feature-development.md # Complete feature implementation
β”œβ”€β”€ bug-investigation.md   # Bug investigation and fixing
└── code-review.md         # Code review and improvement

CLAUDE.md                  # Complete agent system documentation

πŸ€– Available Agents

Agent Purpose When to Use Model
@agent-orchestrator Coordinate multi-agent workflows Complex tasks requiring multiple agents Opus
@agent-architect Design solutions, analyze requirements Planning new features, architectural decisions Opus
@agent-codebase-finder Locate code patterns and implementations Finding existing code, understanding patterns Haiku
@agent-debugger Systematic bug investigation Complex debugging, root cause analysis Opus
@agent-developer Implement code and tests Writing code, implementing features Inherit
@agent-quality-reviewer Quality validation and review Code review, quality assessment Sonnet
@agent-technical-writer Create documentation After feature completion Sonnet
@agent-feature-enhancer Enhance existing features with new capabilities Improving existing features, adding functionality Opus

🎯 Common Use Cases

Quick Commands (Recommended)

/orchestrate feature auth email/password with Firebase
/orchestrate enhance dashboard with analytics capabilities
/orchestrate bug memory leak in user sessions
/orchestrate review user authentication security focus
/orchestrate docs api user-service endpoints
/orchestrate arch design user authentication system

Full Orchestrator (For Complex Tasks)

@agent-orchestrator: Create a real-time chat feature with WebSocket connections, message persistence, typing indicators, and mobile-responsive design.
@agent-orchestrator: Debug and fix the memory leak in user session management. The app crashes after 100+ logins with increasing memory usage.
@agent-orchestrator: Perform comprehensive review of the authentication system. Focus on security vulnerabilities, TypeScript compliance, and code quality.
@agent-orchestrator: Optimize dashboard loading performance from 5+ seconds to under 2 seconds. Investigate bottlenecks and implement improvements.
@agent-orchestrator: Investigate and fix the production issue where 15% of users are getting randomly logged out. Coordinate debugging, analysis, and deployment.

Feature Enhancement (For Improving Existing Features)

@agent-feature-enhancer: Enhance the user dashboard with advanced analytics capabilities. Add real-time data visualization, custom reporting, and export functionality.
@agent-feature-enhancer: Improve the authentication system with social login options, two-factor authentication, and account recovery features.
@agent-feature-enhancer: Add mobile optimization to the existing data table component with touch gestures, responsive columns, and offline support.
@agent-feature-enhancer: Enhance the search functionality with filters, sorting, saved searches, and search history.
@agent-feature-enhancer: Add collaboration features to the existing document editor with real-time editing, comments, and version control.

πŸ”§ How /orchestrate Works

The /orchestrate command combines high-level workflow coordination with detailed operational guidance:

Built-in Workflow Intelligence

The orchestrator automatically selects and coordinates agent sequences:

  • feature β†’ Architect β†’ Codebase Finder β†’ Developer β†’ Quality Reviewer β†’ Technical Writer
  • enhance β†’ Codebase Finder β†’ Feature Enhancer β†’ Architect β†’ Developer β†’ Quality Reviewer β†’ Technical Writer
  • bug β†’ Debugger β†’ Architect (if needed) β†’ Developer β†’ Quality Reviewer
  • review β†’ Quality Reviewer β†’ Developer (if fixes needed)
  • docs β†’ Technical Writer (with context from other agents)
  • arch β†’ Architect β†’ Quality Reviewer β†’ Technical Writer

Operational Guidance from /workflows

The orchestrator leverages detailed workflow templates that provide:

  • Specific context requirements between agents
  • Detailed quality validation criteria
  • Actionable success criteria with checkboxes
  • Specialized techniques for different scenarios
  • Advanced review methodologies
  • Workflow variations for different complexity levels

Smart Defaults & Context

  • Technology stack detection (Next.js, Firebase, etc.)
  • Quality standards from your project configuration
  • Architecture patterns recognition
  • Error handling following project patterns

πŸ“‹ Best Practices

βœ… Good Orchestrator Calls:

  • Be specific about what you want to accomplish
  • Include context about the current state or requirements
  • Mention quality standards you want to maintain
  • Specify the scope (new feature, bug fix, improvement, etc.)
  • Include any constraints or requirements

❌ Avoid These:

  • Vague requests like "fix the code"
  • Single-agent tasks that don't need orchestration
  • Requests without context or requirements
  • Tasks that are too broad without specific goals

🎯 Command Selection Decision Tree:

Need a feature implemented? β†’ @quick-feature
Need to debug/fix an issue? β†’ @quick-bug
Need code review? β†’ @quick-review
Need documentation? β†’ @quick-docs
Need architectural decision? β†’ @quick-arch
Need to enhance existing features? β†’ @agent-feature-enhancer
Complex multi-step task? β†’ @agent-orchestrator
Specific agent task? β†’ @agent-[name]

πŸš€ Quick Commands Benefits

Using Quick Commands provides:

  1. One-Liner Simplicity - Easy to remember and type
  2. Smart Defaults - P 7FD7 re-configured for common scenarios
  3. Context Awareness - Adapts to your project and patterns
  4. Automatic Workflows - No manual agent coordination needed
  5. Quality Assurance - Built-in validation and testing
  6. Consistent Results - Standardized processes and outcomes

πŸ”§ Full Orchestrator Benefits

Using @agent-orchestrator for complex tasks provides:

  1. Custom Workflows - Tailored to specific requirements
  2. Context Preservation - Information flows seamlessly between agents
  3. Error Recovery - Automatic handling of agent failures
  4. Progress Monitoring - Clear visibility into workflow progress
  5. Quality Gates - Validation at each workflow step
  6. Flexible Coordination - Adapt to unique project needs

πŸ“š Documentation

  • CLAUDE.md - Complete agent system documentation with detailed usage guidelines
  • /commands/ - Quick command shortcuts for easy usage
  • /workflows/ - Detailed workflow templates with operational guidance and specialized techniques
  • prompt-engineering.md - Advanced prompt optimization techniques

πŸ”§ Project-Specific Standards

The system includes built-in support for:

  • Next.js 15 with App Router
  • TypeScript with strict mode
  • Firebase/Firestore integration
  • TailwindCSS styling
  • Zustand state management
  • Quality standards and linting requirements

🀝 Acknowledgments

prompt-engineering.md is extracted from Southbridge Research's exceptional analysis of Claude Code prompts. This document has been invaluable for understanding and applying advanced prompt engineering techniques.

When optimizing prompts with these techniques:

Apply ALL prompt engineering patterns from `prompt-engineering.md` to this modification. For EACH change, specify EXACTLY which technique(s) you used:
- Pattern name (e.g., "Progressive Disclosure", "Emphasis Hierarchy")
- Why this pattern applies here
- Expected behavioral impact

CRITICAL: Changes without pattern attribution = task incomplete.

πŸŽ‰ Getting Started

  1. Start with Quick Commands - Use /orchestrate for most tasks
  2. Read /commands/orchestrate.md for detailed command usage
  3. Use @agent-orchestrator for complex, multi-step tasks
  4. Check CLAUDE.md for complete system documentation
  5. Review /workflows for detailed operational guidance
  6. Follow best practices for optimal results

The system is designed to handle everything from simple bug fixes to complete feature implementations with automated quality assurance and documentation.

Quick Start Examples

# Implement a feature
@quick-feature: auth email/password with Firebase

# Debug an issue
@quick-bug: memory leak in user sessions

# Review code
@quick-review: feature user authentication --focus=security

# Generate docs
@quick-docs: api user-service endpoints

# Make architectural decisions
@quick-arch: design user authentication system

About

My claude code customizations / config.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

0