Skip to content

Agent_Selection_Guide

Jump to bottom
Rafal Lagowski edited this page Sep 14, 2025 · 1 revision

Agent Selection Guide

ClaudeAutoPM provides over 35 specialized AI agents to assist with different aspects of software development. This guide helps you choose the right agent for your task based on technology stack, use case, and project requirements.

Quick Selection Matrix

By Technology Stack

Technology Recommended Agent Use Case
React @react-ui-expert React components, hooks, state management
Python Backend @python-backend-expert FastAPI, Django, Flask development
Docker @docker-containerization-expert Containerization, Docker Compose
Kubernetes @kubernetes-orchestrator K8s deployments, Helm charts
AWS @aws-cloud-architect AWS services, infrastructure
Azure @azure-cloud-architect Azure resources, ARM templates
PostgreSQL @postgresql-expert Database design, queries, optimization
MongoDB @mongodb-expert NoSQL schemas, aggregations
Testing @e2e-test-engineer End-to-end testing, test automation

By Task Type

Task Agent Description
Code Review @code-analyzer Find bugs, security issues, optimizations
Create Tests @test-runner Generate and execute test suites
Build UI @react-ui-expert Create React components with modern patterns
API Development @python-backend-expert Design and implement REST/GraphQL APIs
DevOps Setup @github-operations-specialist CI/CD pipelines, GitHub Actions
Database Design @postgresql-expert Schema design, migrations, optimization
Cloud Architecture @aws-cloud-architect Design scalable cloud solutions
Container Setup @docker-containerization-expert Dockerize applications
Security Audit @code-analyzer Security vulnerability scanning

Core Agents

agent-manager

Location: .claude/agents/core/agent-manager.md Purpose: Manage agent lifecycle and creation When to use:

  • Creating new custom agents
  • Analyzing agent performance
  • Updating agent registry
@agent-manager create a new agent for GraphQL development

code-analyzer

Location: .claude/agents/core/code-analyzer.md Purpose: Analyze code for issues and improvements When to use:

  • Pre-commit code review
  • Security vulnerability scanning
  • Performance optimization analysis
  • Bug investigation
@code-analyzer review this file for security issues and performance problems

test-runner

Location: .claude/agents/core/test-runner.md Purpose: Execute and analyze test results When to use:

  • Running test suites
  • Analyzing test failures
  • Generating test reports
  • Creating new tests
@test-runner execute all tests and provide detailed failure analysis

file-analyzer

Location: .claude/agents/core/file-analyzer.md Purpose: Summarize large files to reduce context When to use:

  • Analyzing large log files
  • Summarizing documentation
  • Processing test outputs
  • Reviewing configuration files
@file-analyzer summarize the key issues in this 5000-line log file

Frontend Development Agents

react-ui-expert

Purpose: React development with modern patterns Expertise:

  • React 18+ features
  • Hooks and custom hooks
  • State management (Redux, Zustand, Context)
  • Component patterns
  • Performance optimization
@react-ui-expert build a data table component with sorting and filtering

javascript-frontend-engineer

Purpose: General JavaScript frontend development Expertise:

  • Vanilla JavaScript
  • DOM manipulation
  • Browser APIs
  • Module systems
  • Build tools
@javascript-frontend-engineer optimize this JavaScript for performance

Backend Development Agents

python-backend-expert

Purpose: Python backend development Expertise:

  • FastAPI, Django, Flask
  • Async programming
  • Database integration
  • REST/GraphQL APIs
  • Testing with pytest
@python-backend-expert create a FastAPI endpoint with authentication

nodejs-backend-engineer

Purpose: Node.js backend development Expertise:

  • Express, NestJS, Koa
  • Middleware patterns
  • Database integration
  • Real-time features
  • Microservices
@nodejs-backend-engineer implement WebSocket server with Socket.io

Database Agents

postgresql-expert

Purpose: PostgreSQL database management Expertise:

  • Schema design
  • Query optimization
  • Indexes and performance
  • Migrations
  • Advanced features (JSONB, full-text search)
@postgresql-expert optimize this slow query with proper indexes

mongodb-expert

Purpose: MongoDB NoSQL development Expertise:

  • Document design
  • Aggregation pipelines
  • Indexing strategies
  • Sharding
  • Performance tuning
@mongodb-expert design schema for user activity tracking

DevOps Agents

docker-containerization-expert

Purpose: Docker and containerization Expertise:

  • Dockerfile optimization
  • Docker Compose
  • Multi-stage builds
  • Security best practices
  • Volume management
@docker-containerization-expert create optimized Dockerfile for Python app

kubernetes-orchestrator

Purpose: Kubernetes deployment and management Expertise:

  • Deployment configurations
  • Service mesh
  • Helm charts
  • Scaling strategies
  • Security policies
@kubernetes-orchestrator create Helm chart for microservices deployment

github-operations-specialist

Purpose: GitHub Actions and CI/CD Expertise:

  • Workflow creation
  • Actions marketplace
  • Release automation
  • Security scanning
  • Matrix testing
@github-operations-specialist create CI/CD pipeline with testing and deployment

Cloud Architecture Agents

aws-cloud-architect

Purpose: AWS infrastructure design Expertise:

  • EC2, Lambda, ECS/EKS
  • S3, RDS, DynamoDB
  • VPC, Security Groups
  • CloudFormation, CDK
  • Cost optimization
@aws-cloud-architect design serverless architecture for image processing

azure-cloud-architect

Purpose: Azure infrastructure design Expertise:

  • Azure VMs, AKS, Functions
  • Storage accounts, SQL Database
  • Virtual Networks, NSGs
  • ARM templates, Bicep
  • Azure DevOps integration
@azure-cloud-architect setup AKS cluster with monitoring

gcp-cloud-architect

Purpose: Google Cloud Platform design Expertise:

  • Compute Engine, GKE, Cloud Run
  • Cloud Storage, Cloud SQL
  • VPC, Firewall rules
  • Terraform, Deployment Manager
  • BigQuery, Pub/Sub
@gcp-cloud-architect implement event-driven architecture with Pub/Sub

Testing Agents

e2e-test-engineer

Purpose: End-to-end testing Expertise:

  • Playwright, Cypress, Selenium
  • Test automation frameworks
  • CI/CD integration
  • Cross-browser testing
  • Performance testing
@e2e-test-engineer create Playwright tests for checkout flow

playwright-test-engineer

Purpose: Playwright-specific testing Expertise:

  • Playwright API
  • Page object model
  • Visual testing
  • API testing
  • Mobile testing
@playwright-test-engineer implement visual regression tests

Decision Matrices

Choosing a UI Framework Agent

graph TD
    A[UI Development] --> B{Framework?}
    B -->|React| C[@react-ui-expert]
    B -->|Vue| D[@javascript-frontend-engineer]
    B -->|Angular| D
    B -->|Vanilla JS| D
    C --> E{Styling?}
    E -->|Tailwind| F[@tailwindcss-expert]
    E -->|Material UI| G[@mui-react-expert]
    E -->|Ant Design| H[@antd-react-expert]
Loading

Choosing a Backend Agent

graph TD
    A[Backend Development] --> B{Language?}
    B -->|Python| C[@python-backend-expert]
    B -->|Node.js| D[@nodejs-backend-engineer]
    B -->|Go| E[@code-analyzer + Custom]
    C --> F{Framework?}
    F -->|FastAPI| G[Use @python-backend-expert]
    F -->|Django| G
    F -->|Flask| G
    D --> H{Framework?}
    H -->|Express| I[Use @nodejs-backend-engineer]
    H -->|NestJS| I
Loading

Agent Combinations

Full Stack Development

# Frontend + Backend + Database
@react-ui-expert create user dashboard
@python-backend-expert create API endpoints
@postgresql-expert design database schema

DevOps Pipeline

# Containerization + CI/CD + Cloud
@docker-containerization-expert dockerize application
@github-operations-specialist setup GitHub Actions
@kubernetes-orchestrator deploy to K8s

Testing Suite

# Unit + Integration + E2E
@test-runner create unit tests
@code-analyzer verify test coverage
@e2e-test-engineer create end-to-end tests

Custom Agent Creation

When existing agents don't meet your needs:

@agent-manager create a custom agent for [specific technology/task]

# Example:
@agent-manager create a GraphQL specialist agent with:
- Apollo Server expertise
- Schema design patterns
- Resolver optimization
- Subscription handling

Best Practices

1. Start Specific

Choose the most specific agent for your task:

  • ❌ @code-analyzer for React component
  • ✅ @react-ui-expert for React component

2. Combine Agents

Use multiple agents for complex tasks:

@code-analyzer review for issues
@test-runner verify with tests
@docker-containerization-expert prepare for deployment

3. Context Matters

Provide context to agents:

# Good
@postgresql-expert optimize query for table with 10M rows

# Better
@postgresql-expert optimize this e-commerce order query:
- Orders table: 10M rows
- Needs to join with users and products
- Used for real-time dashboard

4. Iterative Refinement

Work iteratively with agents:

1. @python-backend-expert create initial API
2. @code-analyzer review for improvements
3. @test-runner add test coverage
4. @docker-containerization-expert containerize

Performance Considerations

Agent Efficiency Ranking

Agent Context Usage Speed Best For
@file-analyzer Low Fast Large files
@code-analyzer Medium Medium Code review
@react-ui-expert Medium Fast React development
@kubernetes-orchestrator High Slow Complex deployments
@aws-cloud-architect High Slow Infrastructure design

Parallel vs Sequential

Parallel Execution (when enabled):

# These run simultaneously
@code-analyzer review backend/
@test-runner execute tests
@docker-containerization-expert build images

Sequential Execution:

# These run one after another
@code-analyzer review and fix issues
THEN @test-runner verify fixes
THEN @docker-containerization-expert deploy

Troubleshooting

Agent Not Responding

# Check if agent exists
ls .claude/agents/ | grep agent-name

# Try core agent instead
@code-analyzer [fallback for specialized agent]

Wrong Agent Selected

# Be more specific
@python-backend-expert for FastAPI development
NOT @code-analyzer for FastAPI development

Agent Limitations

Some agents have specific limitations:

  • Cloud agents need credentials configured
  • Test agents need test framework installed
  • DevOps agents need tools (Docker, kubectl)

Related Pages

Clone this wiki locally