Agents Full Reference

Complete reference for all 40 agents available in Claude Craft.

---

Quick Reference

Category	Count	Examples
Common	13	@api-designer, @tdd-coach, @uiux-orchestrator
Technology Reviewers	10	@symfony-reviewer, @react-reviewer
Docker/Infrastructure	5	@docker-architect, @docker-debug
BMAD v6	10	@bmad-master, @pm, @dev, @qa-recette
Project	2	@product-owner, @tech-lead

---

How to Use Agents

@agent-name Your request here

Multiple Agents

@tdd-coach @react-reviewer Help me TDD this component

With Commands

@dev Implement the following:
/sprint:next-story --claim

---

Common Agents (13)

@api-designer

Expertise: REST/GraphQL API design

Use when:

• Designing new APIs
• API versioning decisions
• Request/response schemas
• OpenAPI documentation

Example prompts:

@api-designer Design REST endpoints for user management
@api-designer Should this be REST or GraphQL?
@api-designer Create OpenAPI spec for the orders API

---

@database-architect

Expertise: Database design and optimization

Use when:

• Schema design
• Query optimization
• Index strategies
• Migration planning

Example prompts:

@database-architect Design schema for e-commerce orders
@database-architect Optimize this slow query
@database-architect Plan migration for splitting the users table

---

@devops-engineer

Expertise: CI/CD, Docker, deployment

Use when:

• Setting up pipelines
• Docker configuration
• Cloud infrastructure
• Monitoring setup

Example prompts:

@devops-engineer Set up GitHub Actions for this project
@devops-engineer Optimize this Dockerfile
@devops-engineer Design monitoring strategy

---

@performance-auditor

Expertise: Performance analysis and optimization

Use when:

• Diagnosing slowness
• Memory optimization
• Caching strategies
• Load testing

Example prompts:

@performance-auditor This page loads slowly, help diagnose
@performance-auditor Review this code for performance issues
@performance-auditor Design caching strategy for API responses

---

@refactoring-specialist

Expertise: Safe code refactoring

Use when:

• Code smells detected
• Legacy modernization
• Design pattern application
• Technical debt reduction

Example prompts:

@refactoring-specialist How to refactor this god class?
@refactoring-specialist Plan to modernize this legacy module
@refactoring-specialist Apply repository pattern here

---

@research-assistant

Expertise: Technical research and documentation

Use when:

• Library comparisons
• Technology evaluation
• Best practices research
• Learning resources

Example prompts:

@research-assistant Compare Zustand vs Redux for this use case
@research-assistant What's the best testing library for React?
@research-assistant Research authentication options for mobile apps

---

@tdd-coach

Expertise: Test-Driven Development

Use when:

• Learning TDD
• Writing tests first
• Bug fixing with tests
• Testing strategies

Example prompts:

@tdd-coach Guide me through TDD for this feature
@tdd-coach Help me write a failing test for this bug
@tdd-coach What testing strategy for this module?

---

@uiux-orchestrator

Expertise: UI/UX coordination

Use when:

• Need combined UI + UX + A11y expertise
• Comprehensive interface design
• Coordinating design decisions

Example prompts:

@uiux-orchestrator Design a complete login flow
@uiux-orchestrator Review this component for UI, UX, and accessibility

---

@ui-designer

Expertise: Design systems and visual design

Use when:

• Design tokens
• Component specifications
• Grid systems
• Visual consistency

Example prompts:

@ui-designer Create design tokens for a new project
@ui-designer Specify a Button component with all variants
@ui-designer Design the grid system for this app

---

@ux-ergonome

Expertise: User experience and cognitive ergonomics

Use when:

• User flow design
• Form optimization
• Information architecture
• Usability issues

Example prompts:

@ux-ergonome Design the checkout flow for e-commerce
@ux-ergonome Analyze this form for usability issues
@ux-ergonome Optimize the navigation for mobile

---

@accessibility-expert

Expertise: WCAG 2.2 AAA accessibility

Use when:

• Accessibility audits
• ARIA implementation
• Keyboard navigation
• Screen reader compatibility

Example prompts:

@accessibility-expert Audit this page for WCAG 2.2 AAA
@accessibility-expert Specify ARIA attributes for a modal dialog
@accessibility-expert Review color contrast in this design

---

@ralph-conductor

Expertise: Continuous AI loop orchestration

Use when:

• Setting up Ralph sessions
• Configuring Definition of Done
• Managing iteration cycles

Example prompts:

@ralph-conductor Set up a Ralph session with TDD DoD
@ralph-conductor Configure DoD for this feature implementation
@ralph-conductor What's the current Ralph iteration status?

---

Technology Reviewers (10)

@symfony-reviewer

Expertise: Symfony/PHP code review

Focuses on:

• Clean Architecture compliance
• DDD patterns
• Symfony best practices
• Doctrine ORM

Example prompts:

@symfony-reviewer Review this service class
@symfony-reviewer Is this aggregate root correct?
@symfony-reviewer Check this controller for security issues

---

@flutter-reviewer

Expertise: Flutter/Dart code review

Focuses on:

• BLoC/Riverpod patterns
• Widget composition
• Performance
• Platform-specific code

Example prompts:

@flutter-reviewer Review this widget for performance
@flutter-reviewer Is this BLoC implementation correct?
@flutter-reviewer Check state management in this feature

---

@react-reviewer

Expertise: React/TypeScript code review

Focuses on:

• Hooks patterns
• State management
• Accessibility
• Performance

Example prompts:

@react-reviewer Review this custom hook
@react-reviewer Check this component for re-renders
@react-reviewer Audit accessibility of this form

---

@reactnative-reviewer

Expertise: React Native code review

Focuses on:

• Navigation patterns
• Native module integration
• Performance
• App store requirements

Example prompts:

@reactnative-reviewer Review this screen component
@reactnative-reviewer Check this native bridge code
@reactnative-reviewer Optimize this list for performance

---

@python-reviewer

Expertise: Python code review

Focuses on:

• Pythonic patterns
• Type hints
• Async/await
• FastAPI/Django

Example prompts:

@python-reviewer Review this FastAPI endpoint
@python-reviewer Check this async code for issues
@python-reviewer Improve type hints in this module

---

@angular-reviewer

Expertise: Angular code review

Focuses on:

• Signals patterns
• Standalone components
• RxJS usage
• Change detection

Example prompts:

@angular-reviewer Review this standalone component
@angular-reviewer Check RxJS usage for memory leaks
@angular-reviewer Optimize change detection strategy

---

@laravel-reviewer

Expertise: Laravel code review

Focuses on:

• Actions pattern
• Pest PHP testing
• Sanctum auth
• Eloquent optimization

Example prompts:

@laravel-reviewer Review this Action class
@laravel-reviewer Check Eloquent queries for N+1
@laravel-reviewer Audit Sanctum configuration

---

@vuejs-reviewer

Expertise: Vue.js code review

Focuses on:

• Composition API
• Pinia stores
• Vitest testing
• TypeScript

Example prompts:

@vuejs-reviewer Review this composable
@vuejs-reviewer Check Pinia store for best practices
@vuejs-reviewer Audit reactivity in this component

---

@csharp-reviewer

Expertise: C#/.NET code review

Focuses on:

• Clean Architecture
• CQRS patterns
• Entity Framework
• Async/await

Example prompts:

@csharp-reviewer Review this command handler
@csharp-reviewer Check EF Core queries for performance
@csharp-reviewer Audit this API endpoint for security

---

@php-reviewer

Expertise: PHP code review

Focuses on:

• PSR-12 compliance
• PHPStan analysis
• Pest PHP testing
• Clean Architecture

Example prompts:

@php-reviewer Review this use case
@php-reviewer Check for PSR-12 violations
@php-reviewer Audit dependency injection

---

Docker/Infrastructure Agents (5)

@docker-dockerfile

Expertise: Dockerfile optimization

Use when:

• Creating Dockerfiles
• Multi-stage builds
• Image size reduction
• Security hardening

Example prompts:

@docker-dockerfile Optimize this Dockerfile for production
@docker-dockerfile Review my multi-stage build
@docker-dockerfile Reduce image size by 30%

---

@docker-compose

Expertise: Docker Compose orchestration

Use when:

• Multi-service setup
• Network configuration
• Volume management
• Environment configs

Example prompts:

@docker-compose Design compose file for microservices
@docker-compose Configure service dependencies
@docker-compose Set up development vs production overrides

---

@docker-debug

Expertise: Docker troubleshooting

Use when:

• Container won't start
• Network issues
• Permission problems
• Resource exhaustion

Example prompts:

@docker-debug Container keeps restarting, help diagnose
@docker-debug Network connection refused between services
@docker-debug Why is this container running out of memory?

---

@docker-cicd

Expertise: Docker CI/CD pipelines

Use when:

• GitHub Actions setup
• GitLab CI pipelines
• Security scanning
• Deployment strategies

Example prompts:

@docker-cicd Create GitHub Actions pipeline for this app
@docker-cicd Set up automated image scanning
@docker-cicd Implement blue-green deployment

---

@docker-architect

Expertise: Docker architecture design

Use when:

• Complete Docker strategy
• Service topology
• Scaling strategy
• Multi-environment setup

Example prompts:

@docker-architect Design Docker architecture for e-commerce platform
@docker-architect Review our microservices container strategy
@docker-architect Plan scaling strategy for production

---

BMAD v6 Roles (10)

Note: These are personas integrated into workflow and sprint commands, not standalone agent files.

@qa-recette

Expertise: Automated Acceptance Testing with Claude in Chrome

Use when:

• Running acceptance tests (recette) on web applications
• Validating User Stories, Epics, or Sprints
• Detecting regressions after bug fixes
• Generating regression tests from errors

Chrome Capabilities:

• navigate, click, type, fill_form, scroll
• read_console_logs, get_network_requests
• read_dom_state, take_screenshot, record_gif

Key Commands: /qa:recette, /qa:fix, /qa:status, /qa:regression, /qa:report

Golden Rule: A fixed bug should NEVER reappear - automatic regression suite generation.

Example prompts:

@qa-recette Run acceptance tests for user story US-001
@qa-recette Check for regressions on the authentication module
@qa-recette Generate regression tests from the login error

---

@bmad-master

Role: Central orchestrator

Use when:

• Starting BMAD
• Routing work
• Checking status
• Coordinating agents

Key Commands: /bmad:init, /bmad:status, /bmad:route

---

@pm

Role: Product Manager

Use when:

• Creating PRD
• Defining vision
• Prioritizing features
• Setting metrics

Key Commands: /pm:prd, /pm:vision, /pm:roadmap

Quality Gate: PRD Gate (≥80%)

---

@ba

Role: Business Analyst

Use when:

• Gathering requirements
• Creating use cases
• Story mapping
• Gap analysis

Key Commands: /ba:analyze, /ba:use-cases, /ba:story-map

---

@architect

Role: System Architect

Use when:

• Designing architecture
• Creating tech specs
• Writing ADRs
• API design

Key Commands: /arch:design, /arch:techspec, /arch:adr

Quality Gate: Tech Spec Gate (≥90%)

---

@po

Role: Product Owner

Use when:

• Backlog management
• Writing stories
• Acceptance criteria
• Accepting work

Key Commands: /po:prioritize, /po:refine, /po:accept

---

@sm

Role: Scrum Master

Use when:

• Sprint planning
• Running ceremonies
• Tracking velocity
• Removing blockers

Key Commands: /sm:plan-sprint, /sm:daily, /sm:retro

---

@dev

Role: Developer

Use when:

• TDD implementation
• Code review
• Refactoring
• Bug fixing

Key Commands: /dev:implement, /dev:tdd, /dev:refactor

TDD Phases: 🔴 Red → 🟢 Green → 🔵 Refactor

---

@qa

Role: QA Engineer

Use when:

• Test strategy
• Validating AC
• Test automation
• Quality reporting

Key Commands: /qa:strategy, /qa:validate, /qa:automate

---

@ux

Role: UX Designer

Use when:

• User research
• Personas/journeys
• Wireframes
• Accessibility

Key Commands: /ux:research, /ux:journey, /ux:wireframe

---

Project Agents (2)

@product-owner

Expertise: Product management (CSPO)

Use when:

• Story writing
• Backlog prioritization
• Acceptance criteria
• Release planning

---

@tech-lead

Expertise: Technical leadership (CSM)

Use when:

• Task decomposition
• Architecture decisions
• Sprint estimation
• Technical debt

---

Memory Frontmatter (v2.1.33+)

The memory field enables persistent memory for agents across sessions. When enabled, the agent can read and write files in its memory directory, accumulating knowledge over time.

Syntax:

memory: user    # or project, local

Scope Details:

Scope	Directory	In .gitignore	Best For
user	~/.claude/agent-memory/<agent-name>/	N/A (user home)	Personal learnings, cross-project patterns
project	.claude/agent-memory/<agent-name>/	No (shared)	Team knowledge, project conventions
local	.claude/agent-memory-local/<agent-name>/	Yes	Local experiments, draft notes

Example - Research Agent with Memory:

---
name: research-assistant
description: Technical research with persistent knowledge
model: sonnet
memory: user
---

The agent will automatically have access to read/write files in ~/.claude/agent-memory/research-assistant/, allowing it to build up knowledge across sessions.

---

Agent Type Restrictions (v2.1.33+)

The tools frontmatter field now supports Task(agent_type) syntax to restrict which sub-agent types an agent can spawn:

---
name: safe-researcher
tools:
  - Task(Explore)    # Can spawn Explore agents
  - Task(Plan)       # Can spawn Plan agents
  - Read             # Can read files
  - Grep             # Can search content
  - Glob             # Can find files
  # No Task(Bash) or Task(general-purpose) = cannot spawn those
---

Available agent types for restriction:

• Task(Explore) - Read-only exploration agents
• Task(Plan) - Planning agents (read-only)
• Task(Bash) - Command execution agents
• Task(general-purpose) - Full-capability agents

---

Creating Custom Agents

Add to .claude/agents/:

---
name: my-agent
description: Expert in X. Use when Y.
---

# My Agent

## Identity
- **Name**: My Agent
- **Role**: Expert
- **Expertise**: Domain knowledge

## Capabilities
- Capability 1
- Capability 2

## Methodology
How this agent approaches problems...

## Output Format
Expected output structure...

---

Best Practices

1. Be Specific - Clear requests get better results
2. Provide Context - Share relevant code and requirements
3. Use Appropriate Agent - Match expertise to task
4. Combine When Needed - Multiple agents for complex tasks
5. Follow Suggestions - Agents guide, not just answer

---

See Also

• Commands Reference
• BMAD Practical Guide
• Quick Reference