Skip to main content
Everything Claude Code includes 13 specialized agents that handle specific aspects of software development. These agents work autonomously or can be invoked explicitly when needed.

Agent Philosophy

Agent-First Development: Delegate to specialized agents for domain-specific tasks. Agents are designed to be proactive — they automatically activate when their expertise is needed.

Available Agents

Planner

Implementation planning for complex features and refactoring

Architect

System design, scalability, and technical decision-making

TDD Guide

Test-driven development with 80%+ coverage enforcement

Code Reviewer

Code quality, security, and maintainability review

Build Error Resolver

Fix TypeScript and build errors quickly

Security Reviewer

Vulnerability detection and OWASP compliance

E2E Runner

End-to-end testing with Playwright

Refactor Cleaner

Dead code cleanup and consolidation

Doc Updater

Documentation and codemap generation

Database Reviewer

PostgreSQL optimization and schema design

Go Reviewer

Idiomatic Go code review

Go Build Resolver

Go compilation error resolution

Python Reviewer

Pythonic patterns and PEP 8 compliance

Agent Summary Table

Agent Orchestration

Automatic Activation

Agents activate proactively based on context:
  • Complex feature request → planner
  • Code written/modified → code-reviewer
  • Bug fix or new feature → tdd-guide
  • Architectural decision → architect
  • Security-sensitive code → security-reviewer
  • Build failure → build-error-resolver or go-build-resolver
  • Database query/migration → database-reviewer

Parallel Execution

For independent operations, multiple agents can run simultaneously:

When to Use Each Agent

  • planner: Breaking down large features, refactoring plans, risk analysis
  • architect: System design, technology choices, scalability planning
  • tdd-guide: Writing tests first, ensuring coverage
  • e2e-runner: Critical user flow testing
  • code-reviewer: General code review after writing code
  • go-reviewer: Go-specific review
  • python-reviewer: Python-specific review
  • database-reviewer: SQL query and schema review
  • build-error-resolver: TypeScript/JavaScript build errors
  • go-build-resolver: Go compilation errors
  • refactor-cleaner: Dead code removal
  • doc-updater: Keep documentation current
  • security-reviewer: Regular security audits

Best Practices

Use agents proactively — don’t wait for user prompts
Run independent agents in parallel for better performance
Trust agent specialization — let them handle their domain
Review agent recommendations before applying fixes
Agents make minimal changes by design — they won’t refactor unless specifically asked

Next Steps

Learn About Planner

Deep dive into implementation planning

TDD Workflow

Master test-driven development

Security Best Practices

Prevent vulnerabilities before deployment

Database Optimization

Write performant PostgreSQL queries