Claude Code subagent imported from qdhenry/Claude-Command-Suite (.claude/agents/README.md). Copyright stays with the author.

🤖 Claude Command Suite AI Agents

Agents Workflows Context Isolation Automation

Transform your development workflow with intelligent AI agents that work proactively to maintain code quality, automate tasks, and accelerate delivery.

Quick Start • Available Agents • Workflows • Examples • Best Practices

🌟 What Are Claude Code Agents?

Claude Command Suite Agents are specialized AI assistants that transform the 110+ slash commands into an intelligent, proactive development ecosystem. Each agent:

• 🧠 Has Specialized Expertise: Focused on specific domains (security, testing, architecture, etc.)
• 🔒 Operates in Isolation: Separate context windows prevent cross-contamination
• ⚡ Works Proactively: Automatically triggered based on context and keywords
• 🔧 Uses Minimal Tools: Only granted necessary permissions for efficiency
• 🔄 Chains Together: Combine agents for complex multi-step workflows

Why Use Agents?

🚀 Quick Start

1. Basic Usage

# Explicit invocation
"Use the code-auditor agent to review my changes"

# Automatic triggering  
"I need to fix the security vulnerability in the login system"
# → Automatically activates security-auditor

# Agent chaining
"First analyze our architecture, then create tests for the new module"
# → architecture-auditor → test-engineer

2. Common Patterns

# Quality Assurance Flow
"Review this PR thoroughly"
# Activates: code-auditor → security-auditor → test-engineer

# New Feature Development
"Set up a new user dashboard feature"  
# Activates: project-architect → test-engineer → integration-manager

# Release Preparation
"Prepare version 2.0 for release"
# Activates: All auditors → release-manager

🤖 Available Agents

🛡️ Code Quality Assurance Suite

Agent: code-auditor
Expertise: Code quality, patterns, best practices
Auto-triggers: After code changes, PR reviews, refactoring
Tools: Read, Grep, Glob, Bash, WebFetch

# Example usage
"Review the payment module for code quality issues"
"Check if our API follows REST best practices"

Agent: security-auditor
Expertise: OWASP Top 10, security vulnerabilities, threat modeling
Auto-triggers: Authentication code, API endpoints, data handling
Tools: Read, Grep, Glob, WebFetch

# Example usage
"Check for security vulnerabilities in the user authentication"
"Review our API for potential injection attacks"

Agent: performance-auditor
Expertise: Performance bottlenecks, optimization, profiling
Auto-triggers: Slow queries, large datasets, performance issues
Tools: Read, Grep, Glob, Bash

# Example usage
"Find performance bottlenecks in the data processing module"
"Optimize the dashboard loading time"

Agent: architecture-auditor
Expertise: Design patterns, SOLID principles, system architecture
Auto-triggers: New features, refactoring, architecture decisions
Tools: Read, Grep, Glob, Bash

# Example usage
"Review our microservice boundaries"
"Check if the codebase follows clean architecture principles"

🔧 Development & Engineering

Agent: test-engineer
Expertise: Test generation, coverage analysis, TDD/BDD
Auto-triggers: New code, low coverage, bug fixes
Tools: Read, Write, Edit, Bash, Grep, Glob

# Example usage
"Generate comprehensive tests for the UserService class"
"Create integration tests for the payment flow"

Agent: project-architect
Expertise: Project scaffolding, best practices, tooling setup
Auto-triggers: New projects, major features, restructuring
Tools: Read, Write, Edit, Bash, Glob, TodoWrite

# Example usage
"Set up a new React TypeScript project with best practices"
"Create a microservice structure for inventory management"

🔄 Operations & Integration

Agent: integration-manager
Expertise: GitHub-Linear sync, API integration, webhooks
Auto-triggers: Issue updates, PR events, sync requests
Tools: Bash, Read, Write, MCP Linear tools, WebFetch

# Example usage
"Sync all GitHub issues to Linear"
"Set up bidirectional sync between GitHub and Linear"

Agent: release-manager
Expertise: Versioning, changelogs, deployment, rollbacks
Auto-triggers: Version tags, release prep, hotfixes
Tools: Read, Write, Edit, Bash, Grep, Glob, WebFetch

# Example usage
"Prepare version 2.1.0 for release"
"Create a hotfix release for the critical bug"

📊 Strategic Planning

Agent: strategic-analyst
Expertise: Scenario planning, risk assessment, decision modeling
Auto-triggers: Major decisions, planning sessions, architecture choices
Tools: Read, Write, Bash, WebFetch, MCP Linear tools

# Example usage
"Model scenarios for migrating to microservices"
"Analyze the business impact of our API redesign"

🔄 Agent Workflows

1. Development Workflow

graph LR
    A[Code Change] --> B[code-auditor]
    B --> C{Issues Found?}
    C -->|Yes| D[security-auditor]
    C -->|Yes| E[performance-auditor]
    C -->|No| F[test-engineer]
    D --> F
    E --> F
    F --> G[release-manager]

2. Feature Development Flow

graph TD
    A[New Feature Request] --> B[strategic-analyst]
    B --> C[project-architect]
    C --> D[Development]
    D --> E[test-engineer]
    E --> F[code-auditor Suite]
    F --> G[integration-manager]
    G --> H[release-manager]

3. Bug Fix Workflow

graph TD
    A[Bug Report] --> B[integration-manager]
    B --> C[code-auditor]
    C --> D[Root Cause Analysis]
    D --> E[Fix Implementation]
    E --> F[test-engineer]
    F --> G[security-auditor]
    G --> H[release-manager]

📋 Agent Capabilities Matrix

🎯 Best Practices

1. Agent Selection

• Be Specific: Clear requests help agents understand their task
• Chain Wisely: Combine agents for complex workflows
• Trust Automation: Let agents work proactively

2. Workflow Optimization

# Parallel Execution (Faster)
"Have all auditors review the codebase simultaneously"

# Sequential Execution (Dependencies)
"First run security-auditor, then fix issues, then test-engineer"

3. Context Management

• Agents preserve main conversation context
• Each agent starts fresh for focused analysis
• Results are summarized back to main thread

📊 Performance Metrics

Track your agent usage effectiveness:

🔧 Creating Custom Agents

Template Structure

---
name: your-agent-name
description: Clear description of when to use this agent. MUST BE USED for [specific triggers]. Use PROACTIVELY for [scenarios].
tools: Tool1, Tool2, Tool3  # Minimal necessary tools
---

You are an expert in [domain]. Your role is to [specific responsibilities].

## Core Expertise
- [Expertise area 1]
- [Expertise area 2]

## Working Process
1. [Step 1]
2. [Step 2]

## Output Format
[Define structured output]

Implementation Tips

1. Single Responsibility: One agent, one domain
2. Clear Triggers: Use "MUST BE USED" and "PROACTIVELY"
3. Minimal Tools: Only what's necessary
4. Structured Output: Consistent formatting

🌐 External Agents

wshobson's Agent Collection

We've integrated 44 specialized agents from wshobson/agents, providing additional expertise across:

• Development Specialists: backend-architect, frontend-developer, mobile-developer, swift-macos-expert
• Language Experts: python-pro, golang-pro, rust-pro, javascript-pro, and more
• Infrastructure & DevOps: devops-troubleshooter, cloud-architect, terraform-specialist
• Database Specialists: database-admin, database-optimizer, sql-pro
• Quality & Security: code-reviewer, security-auditor, performance-engineer

These agents are located in .claude/agents/external/wshobson/ and can be used exactly like native agents:

# Examples
"Use backend-architect to design the API structure"
"Have security-auditor check for OWASP compliance"
"Get python-pro to optimize this data processing script"

See ATTRIBUTION.md for full attribution details.

🚦 Troubleshooting

• Verify agent file exists in .claude/agents/ or .claude/agents/external/
• Check YAML frontmatter formatting
• Ensure description contains trigger keywords
• Try explicit invocation first

• Check tool permissions match agent needs
• Verify no circular agent dependencies
• Review system prompt for clarity
• Check MCP server connections

• Reduce tool access to minimum
• Simplify agent prompts
• Break complex tasks into steps
• Use parallel execution where possible

🔮 Roadmap

Coming Soon

• Agent Communication: Direct agent-to-agent messaging
• Shared Memory: Persistent context between invocations
• Learning System: Agents improve from usage patterns
• Custom Templates: Easy agent creation wizards
• Analytics Dashboard: Track agent performance metrics

Future Vision

• Self-Organizing Workflows: Agents automatically form optimal chains
• Predictive Activation: Agents anticipate needs before requests
• Cross-Project Learning: Agents share insights across codebases

📚 Resources

Documentation

• Workflow Examples - Detailed workflow patterns
• Agent Analysis - Deep dive into agent design
• Claude Code Docs - Official documentation

Quick Reference

• List agents: ls .claude/agents/*.md
• Test agent: "Use [agent-name] to [task]"
• Chain agents: "First [agent1] then [agent2]"

🚀 Ready to 10x your development workflow?
Start with "Review my codebase with all quality agents"

Powered by Claude Command Suite - Where commands become intelligent assistants