ruflo/docs/reference/AGENTS.md

🤖 Claude-Flow Agent Reference

Complete Guide to All 65+ Specialized AI Agents

Claude-Flow provides a comprehensive ecosystem of specialized AI agents designed for enterprise-grade software development, coordination, and automation. Each agent is optimized for specific tasks and domains, enabling intelligent swarm coordination and autonomous workflow execution.

---

📊 Agent Overview

Category	Count	Description
Core Development	5	Essential development agents (coder, planner, researcher, reviewer, tester)
Swarm Coordination	5	Swarm topology and coordination management
Consensus & Fault Tolerance	7	Distributed consensus and Byzantine fault tolerance
GitHub Integration	12	Complete GitHub workflow automation
Specialized Domain	8	Domain-specific development (mobile, ML, backend)
Analysis & Code Quality	6	Code analysis, review, and quality assurance
Testing & Validation	4	Comprehensive testing and validation
Infrastructure & DevOps	5	CI/CD, deployment, and infrastructure
SPARC Methodology	4	Specification, Pseudocode, Architecture, Refinement, Code
Template & Automation	9+	Template generation and automation agents

Total Agents: 65+

---

🏗️ 1. Core Development Agents

coder

Type: Implementation Specialist
Purpose: Code generation, refactoring, and implementation
Key Capabilities:

• TypeScript/JavaScript code generation
• API implementation and integration
• Database schema design and implementation
• Code refactoring and optimization
• Bug fixing and feature implementation

Usage Examples:

npx claude-flow@alpha agent spawn coder --name "API-Builder"
npx claude-flow@alpha task assign coder "implement REST API endpoints"

planner

Type: Strategic Planning
Purpose: Project planning, task breakdown, and roadmap creation
Key Capabilities:

• Strategic project planning
• Task decomposition and prioritization
• Timeline estimation and milestone planning
• Resource allocation planning
• Risk assessment and mitigation planning

Usage Examples:

npx claude-flow@alpha agent spawn planner --name "Project-Strategist"
npx claude-flow@alpha sparc run planner "create microservices architecture plan"

researcher

Type: Information Gathering
Purpose: Research, analysis, and knowledge discovery
Key Capabilities:

• Technology research and evaluation
• Best practices discovery
• Market and competitor analysis
• Documentation research
• Problem space exploration

Usage Examples:

npx claude-flow@alpha agent spawn researcher --name "Tech-Scout"
npx claude-flow@alpha task assign researcher "research GraphQL best practices"

reviewer

Type: Quality Assurance
Purpose: Code reviews, best practices enforcement
Key Capabilities:

• Code quality assessment
• Security vulnerability detection
• Best practices enforcement
• Architecture review
• Performance optimization recommendations

Usage Examples:

npx claude-flow@alpha agent spawn reviewer --name "Quality-Guardian"
npx claude-flow@alpha task assign reviewer "review authentication implementation"

tester

Type: Test Creation
Purpose: Unit tests, integration tests, and validation
Key Capabilities:

• Unit test generation (Jest, Mocha)
• Integration test implementation
• E2E test automation
• Test coverage analysis
• Performance testing

Usage Examples:

npx claude-flow@alpha agent spawn tester --name "Test-Master"
npx claude-flow@alpha sparc tdd "user authentication system"

---

🐝 2. Swarm Coordination Agents

hierarchical-coordinator

Type: Queen-Led Coordination
Purpose: Centralized command and control structure
Key Capabilities:

• Centralized decision making
• Task delegation and oversight
• Resource allocation management
• Quality control and validation
• Strategic direction setting

Usage Examples:

npx claude-flow@alpha swarm init --topology hierarchical --coordinator queen-coordinator
npx claude-flow@alpha agent spawn hierarchical-coordinator --name "Queen-Genesis"

mesh-coordinator

Type: Peer-to-Peer Coordination
Purpose: Distributed coordination without central authority
Key Capabilities:

• Peer-to-peer task coordination
• Distributed decision making
• Load balancing across agents
• Fault tolerance through redundancy
• Adaptive task redistribution

Usage Examples:

npx claude-flow@alpha swarm init --topology mesh --max-agents 8
npx claude-flow@alpha agent spawn mesh-coordinator --name "Mesh-Alpha"

adaptive-coordinator

Type: Dynamic Topology Management
Purpose: Adaptive coordination based on workload and conditions
Key Capabilities:

• Dynamic topology adjustment
• Real-time load balancing
• Performance-based agent selection
• Automatic scaling decisions
• Context-aware coordination

Usage Examples:

npx claude-flow@alpha swarm init --topology adaptive --auto-scale
npx claude-flow@alpha agent spawn adaptive-coordinator --name "Adaptive-Prime"

collective-intelligence-coordinator

Type: Hive-Mind Coordination
Purpose: Collective intelligence and shared knowledge
Key Capabilities:

• Shared knowledge base management
• Collective decision making
• Distributed learning and adaptation
• Consensus-based planning
• Emergent intelligence coordination

Usage Examples:

npx claude-flow@alpha hive-mind spawn --collective-intelligence
npx claude-flow@alpha agent spawn collective-intelligence-coordinator --name "Hive-Mind"

swarm-memory-manager

Type: Distributed Memory Coordination
Purpose: Persistent memory and knowledge management
Key Capabilities:

• Distributed memory coordination
• Knowledge persistence and retrieval
• Cross-agent information sharing
• Memory consistency management
• Historical data management

Usage Examples:

npx claude-flow@alpha agent spawn swarm-memory-manager --name "Memory-Keeper"
npx claude-flow@alpha memory distributed init --coordinator swarm-memory-manager

---

⚖️ 3. Consensus & Fault Tolerance Agents

byzantine-coordinator

Type: Byzantine Fault Tolerance
Purpose: Fault tolerance in adversarial conditions
Key Capabilities:

• Byzantine fault tolerance implementation
• Malicious agent detection
• Secure consensus protocols
• Network partition handling
• Security threat mitigation

raft-manager

Type: Leader Election
Purpose: Raft consensus algorithm implementation
Key Capabilities:

• Leader election management
• Log replication coordination
• Consensus state management
• Failure detection and recovery
• Cluster membership management

consensus-builder

Type: Decision-Making Coordination
Purpose: Multi-agent consensus and decision making
Key Capabilities:

• Voting protocol coordination
• Quorum management
• Conflict resolution
• Decision aggregation
• Consensus threshold management

quorum-manager

Type: Quorum Management
Purpose: Quorum-based decision making
Key Capabilities:

• Quorum size determination
• Voting coordination
• Majority decision enforcement
• Member availability tracking
• Quorum recovery management

gossip-coordinator

Type: Gossip Protocol Management
Purpose: Information dissemination and coordination
Key Capabilities:

• Gossip protocol implementation
• Information spreading coordination
• Network topology maintenance
• Rumor tracking and verification
• Epidemic-style communication

crdt-synchronizer

Type: Conflict-Free Replicated Data Types
Purpose: Distributed data synchronization
Key Capabilities:

• CRDT implementation and management
• Conflict-free data synchronization
• Eventual consistency coordination
• Merge operation management
• Distributed state reconciliation

security-manager

Type: Security Coordination
Purpose: Security and access control management
Key Capabilities:

• Access control enforcement
• Security policy management
• Threat detection and response
• Encryption key management
• Audit trail maintenance

---

🐙 4. GitHub Integration Agents

github-modes

Type: Comprehensive GitHub Integration
Purpose: Complete GitHub workflow orchestration
Key Capabilities:

• GitHub workflow orchestration
• Multi-repository coordination
• Branch management and strategies
• Webhook handling and automation
• GitHub API integration

pr-manager

Type: Pull Request Management
Purpose: Automated PR management and review
Key Capabilities:

• Automated PR creation and updates
• Multi-reviewer coordination
• Conflict resolution assistance
• Review assignment optimization
• Merge strategy enforcement

code-review-swarm

Type: Multi-Agent Code Review
Purpose: Distributed code review coordination
Key Capabilities:

• Multi-agent review coordination
• Specialized review assignment
• Code quality assessment
• Security vulnerability scanning
• Review consensus building

issue-tracker

Type: Issue Management
Purpose: Intelligent issue tracking and management
Key Capabilities:

• Issue classification and prioritization
• Automated issue assignment
• Progress tracking and reporting
• Issue relationship mapping
• Resolution coordination

release-manager

Type: Release Coordination
Purpose: Release planning and deployment coordination
Key Capabilities:

• Release planning and scheduling
• Changelog generation
• Deployment coordination
• Rollback management
• Version management

repo-architect

Type: Repository Architecture
Purpose: Repository structure and organization
Key Capabilities:

• Repository structure design
• Branching strategy optimization
• Workflow template creation
• CI/CD pipeline design
• Repository governance

project-board-sync

Type: Project Board Management
Purpose: GitHub project board synchronization
Key Capabilities:

• Project board automation
• Card movement coordination
• Progress visualization
• Milestone tracking
• Team coordination

workflow-automation

Type: GitHub Actions Automation
Purpose: GitHub Actions workflow management
Key Capabilities:

• Workflow design and optimization
• Action marketplace integration
• CI/CD pipeline management
• Secret management
• Workflow debugging

sync-coordinator

Type: Multi-Repository Synchronization
Purpose: Cross-repository coordination
Key Capabilities:

• Multi-repo synchronization
• Dependency tracking
• Cross-repo issue linking
• Unified release coordination
• Repository relationship management

swarm-issue

Type: Swarm-Based Issue Resolution
Purpose: Multi-agent issue resolution
Key Capabilities:

• Swarm-based problem solving
• Issue decomposition
• Parallel resolution strategies
• Resource coordination
• Solution integration

swarm-pr

Type: Swarm-Based PR Management
Purpose: Multi-agent PR handling
Key Capabilities:

• Distributed PR review
• Parallel development coordination
• Merge conflict resolution
• Quality assurance coordination
• Integration testing

multi-repo-swarm

Type: Multi-Repository Swarm Coordination
Purpose: Large-scale repository management
Key Capabilities:

• Multi-repository coordination
• Distributed development management
• Cross-repo dependency tracking
• Unified build coordination
• Release synchronization

---

💻 5. Specialized Domain Agents

backend-dev

Type: Server Development
Purpose: API development, databases, and server-side services
Key Capabilities:

• REST/GraphQL API development
• Database design and optimization
• Microservices architecture
• Authentication and authorization
• Server-side optimization

mobile-dev

Type: Mobile Application Development
Purpose: React Native, iOS, and Android development
Key Capabilities:

• React Native development
• Native iOS/Android integration
• Mobile UI/UX optimization
• Performance optimization
• App store deployment

ml-developer

Type: Machine Learning
Purpose: Model training, deployment, and ML pipelines
Key Capabilities:

• Machine learning model development
• Data pipeline creation
• Model training and optimization
• ML deployment strategies
• Performance monitoring

system-architect

Type: High-Level System Design
Purpose: Architecture design and system planning
Key Capabilities:

• System architecture design
• Scalability planning
• Technology stack selection
• Integration pattern design
• Performance architecture

sparc-coder

Type: SPARC TDD Implementation
Purpose: Test-driven development using SPARC methodology
Key Capabilities:

• SPARC methodology implementation
• Test-driven development
• Specification-based coding
• Iterative refinement
• Quality-focused development

production-validator

Type: Production Validation
Purpose: Real-world validation and testing
Key Capabilities:

• Production environment validation
• Real-world testing scenarios
• Performance validation
• Security assessment
• Deployment verification

api-docs

Type: API Documentation
Purpose: OpenAPI and API documentation generation
Key Capabilities:

• OpenAPI specification generation
• API documentation creation
• Interactive documentation
• Code example generation
• Documentation maintenance

cicd-engineer

Type: CI/CD Pipeline Management
Purpose: Continuous integration and deployment
Key Capabilities:

• CI/CD pipeline design
• Build automation
• Deployment strategies
• Pipeline optimization
• Quality gates implementation

---

🔍 6. Analysis & Code Quality Agents

code-analyzer

Type: Code Analysis
Purpose: Static code analysis and quality assessment
Key Capabilities:

• Static code analysis
• Code complexity assessment
• Technical debt identification
• Refactoring recommendations
• Code pattern recognition

perf-analyzer

Type: Performance Analysis
Purpose: Bottleneck identification and optimization
Key Capabilities:

• Performance bottleneck identification
• Resource usage analysis
• Optimization recommendations
• Load testing coordination
• Performance monitoring

performance-benchmarker

Type: Performance Testing
Purpose: Comprehensive performance benchmarking
Key Capabilities:

• Performance benchmark creation
• Load testing execution
• Stress testing coordination
• Performance regression detection
• Benchmark reporting

analyze-code-quality

Type: Code Quality Assessment
Purpose: Comprehensive code quality analysis
Key Capabilities:

• Code quality metrics
• Maintainability assessment
• Design pattern analysis
• Code smell detection
• Quality improvement recommendations

refactoring-specialist

Type: Code Refactoring
Purpose: Code refactoring and improvement
Key Capabilities:

• Code refactoring strategies
• Design pattern implementation
• Legacy code modernization
• Performance optimization
• Architecture improvement

security-analyzer

Type: Security Analysis
Purpose: Security vulnerability assessment
Key Capabilities:

• Security vulnerability scanning
• Threat modeling
• Security best practices enforcement
• Compliance assessment
• Security remediation

---

🧪 7. Testing & Validation Agents

tdd-london-swarm

Type: London-Style TDD
Purpose: London school test-driven development
Key Capabilities:

• Outside-in TDD approach
• Mock-based testing
• Behavior specification
• Test isolation
• Design emergence

unit-test-specialist

Type: Unit Testing
Purpose: Comprehensive unit test creation
Key Capabilities:

• Unit test generation
• Test coverage optimization
• Test maintainability
• Assertion strategies
• Test performance

integration-tester

Type: Integration Testing
Purpose: System integration validation
Key Capabilities:

• Integration test design
• API testing
• Database integration testing
• Service integration validation
• End-to-end testing

e2e-automation

Type: End-to-End Testing
Purpose: Full system validation
Key Capabilities:

• E2E test automation
• User journey testing
• Cross-browser testing
• Mobile testing
• Visual regression testing

---

🏗️ 8. Infrastructure & DevOps Agents

ops-cicd-github

Type: GitHub CI/CD Operations
Purpose: GitHub Actions and CI/CD management
Key Capabilities:

• GitHub Actions workflow creation
• CI/CD pipeline optimization
• Deployment automation
• Environment management
• Release automation

infrastructure-specialist

Type: Infrastructure Management
Purpose: Cloud infrastructure and deployment
Key Capabilities:

• Infrastructure as Code
• Cloud resource management
• Container orchestration
• Networking configuration
• Monitoring setup

deployment-coordinator

Type: Deployment Management
Purpose: Application deployment coordination
Key Capabilities:

• Deployment strategy implementation
• Blue-green deployments
• Canary releases
• Rollback management
• Environment promotion

monitoring-specialist

Type: System Monitoring
Purpose: Application and infrastructure monitoring
Key Capabilities:

• Monitoring setup and configuration
• Alert management
• Performance tracking
• Log aggregation
• Dashboard creation

cloud-architect

Type: Cloud Architecture
Purpose: Cloud-native architecture design
Key Capabilities:

• Cloud architecture design
• Multi-cloud strategies
• Cost optimization
• Security architecture
• Disaster recovery planning

---

📋 9. SPARC Methodology Agents

specification

Type: SPARC Specification Phase
Purpose: Requirements specification and analysis
Key Capabilities:

• Requirements gathering and analysis
• User story creation
• Acceptance criteria definition
• Specification documentation
• Stakeholder communication

pseudocode

Type: SPARC Pseudocode Phase
Purpose: Algorithm design and pseudocode creation
Key Capabilities:

• Algorithm design
• Pseudocode generation
• Logic flow documentation
• Complexity analysis
• Implementation planning

architecture

Type: SPARC Architecture Phase
Purpose: System architecture and design
Key Capabilities:

• System architecture design
• Component interaction design
• Interface definition
• Technology selection
• Architecture documentation

refinement

Type: SPARC Refinement Phase
Purpose: Design refinement and optimization
Key Capabilities:

• Design refinement
• Performance optimization
• Security enhancement
• Scalability improvement
• Quality validation

---

🔧 10. Template & Automation Agents

base-template-generator

Type: Template Generation
Purpose: Base template and scaffold creation
Key Capabilities:

• Project template generation
• Scaffold creation
• Boilerplate code generation
• Configuration template creation
• Documentation templates

automation-smart-agent

Type: Smart Automation
Purpose: Intelligent automation and workflow creation
Key Capabilities:

• Workflow automation
• Task automation
• Process optimization
• Rule-based automation
• Intelligent scheduling

coordinator-swarm-init

Type: Swarm Initialization
Purpose: Swarm setup and initialization
Key Capabilities:

• Swarm topology setup
• Agent configuration
• Communication channel setup
• Resource allocation
• Initial task distribution

implementer-sparc-coder

Type: SPARC Implementation
Purpose: SPARC-based code implementation
Key Capabilities:

• SPARC methodology implementation
• Code generation from specifications
• Iterative development
• Quality assurance
• Documentation generation

memory-coordinator

Type: Memory Management
Purpose: Distributed memory coordination
Key Capabilities:

• Memory pool management
• Data consistency coordination
• Cache management
• Persistence strategies
• Memory optimization

orchestrator-task

Type: Task Orchestration
Purpose: Complex task coordination and management
Key Capabilities:

• Task decomposition
• Dependency management
• Resource scheduling
• Progress tracking
• Result aggregation

performance-analyzer

Type: Performance Analysis
Purpose: System and application performance analysis
Key Capabilities:

• Performance profiling
• Bottleneck identification
• Resource utilization analysis
• Optimization recommendations
• Performance reporting

sparc-coordinator

Type: SPARC Coordination
Purpose: SPARC methodology coordination
Key Capabilities:

• SPARC phase coordination
• Process orchestration
• Quality gate management
• Milestone tracking
• Deliverable coordination

migration-plan

Type: Migration Planning
Purpose: System and data migration planning
Key Capabilities:

• Migration strategy development
• Risk assessment
• Timeline planning
• Resource allocation
• Rollback planning

---

🚀 Usage Patterns

Single Agent Usage

# Spawn individual agents
npx claude-flow@alpha agent spawn coder --name "API-Builder"
npx claude-flow@alpha agent spawn reviewer --name "Code-Guardian"

# Assign tasks to specific agents
npx claude-flow@alpha task assign coder "implement user authentication"
npx claude-flow@alpha task assign tester "create unit tests for auth module"

Swarm Coordination

# Initialize different swarm topologies
npx claude-flow@alpha swarm init --topology hierarchical --max-agents 8
npx claude-flow@alpha swarm init --topology mesh --agents researcher,coder,tester
npx claude-flow@alpha swarm init --topology adaptive --auto-scale

# Batch agent spawning for complex projects
npx claude-flow@alpha swarm spawn \
  --agents system-architect,backend-dev,frontend-dev,tester,reviewer \
  --task "build e-commerce platform"

SPARC Development Workflow

# Full SPARC methodology execution
npx claude-flow@alpha sparc pipeline "user authentication system"
npx claude-flow@alpha sparc run specification "define user management requirements"
npx claude-flow@alpha sparc run architecture "design auth system architecture"
npx claude-flow@alpha sparc run code "implement authentication module"

GitHub Integration

# GitHub workflow automation
npx claude-flow@alpha github pr-manager "review and merge feature branch"
npx claude-flow@alpha github issue-tracker "manage project issues"
npx claude-flow@alpha github release-manager "prepare v2.0.0 release"

---

🔗 Agent Coordination Patterns

Hierarchical Pattern (Queen-Led)

Queen (hierarchical-coordinator)
├── Architect (system-architect)
├── Workers (coder, backend-dev, mobile-dev)
├── Quality (reviewer, tester)
└── Guardian (security-analyzer)

Mesh Pattern (Peer-to-Peer)

All agents coordinate directly:
coder ↔ reviewer ↔ tester ↔ planner
  ↕       ↕       ↕       ↕
researcher ↔ backend-dev ↔ mobile-dev

Adaptive Pattern (Dynamic)

Coordinator (adaptive-coordinator)
├── Dynamic Agent Pool
├── Load Balancer (load-balancer)
├── Performance Monitor (performance-benchmarker)
└── Auto-scaling Logic

---

📊 Agent Selection Guidelines

Project Type → Recommended Agents

Web Application:

• Core: planner, system-architect, backend-dev, coder, tester
• Quality: reviewer, security-analyzer
• DevOps: cicd-engineer, deployment-coordinator

Mobile Application:

• Core: planner, mobile-dev, backend-dev, api-docs
• Quality: tester, performance-benchmarker
• DevOps: release-manager

Machine Learning Project:

• Core: researcher, ml-developer, data-analyst
• Quality: performance-benchmarker, production-validator
• Infrastructure: cloud-architect, monitoring-specialist

Enterprise System:

• Coordination: hierarchical-coordinator, consensus-builder
• Core: system-architect, backend-dev, security-analyzer
• Quality: code-review-swarm, integration-tester
• DevOps: infrastructure-specialist, monitoring-specialist

---

🛠️ Advanced Configuration

Agent Capabilities Matrix

agent_capabilities:
  coder:
    languages: [typescript, javascript, python]
    frameworks: [react, node.js, express]
    tools: [git, npm, docker]
    max_concurrent_tasks: 3
    
  reviewer:
    analysis_types: [security, performance, quality]
    languages: [typescript, javascript]
    max_review_size: 500_lines
    
  tester:
    test_types: [unit, integration, e2e]
    frameworks: [jest, cypress, playwright]
    coverage_threshold: 80

Custom Agent Creation

# .claude/agents/custom/my-specialist.md
---
name: my-specialist
description: Custom specialized agent for specific domain
capabilities:
  - domain-specific-capability
  - custom-tool-integration
tools: [custom-tool, domain-api]
priority: high
---

# Custom Agent Implementation
[Agent prompt and behavior definition]

---

🔍 Monitoring & Analytics

Agent Performance Metrics

# View agent performance
npx claude-flow@alpha metrics agents --detailed
npx claude-flow@alpha performance analyze --agent-type coder
npx claude-flow@alpha swarm status --topology hierarchical

Real-time Monitoring

# Monitor swarm coordination
npx claude-flow@alpha monitor swarm --real-time
npx claude-flow@alpha dashboard --agents --performance

---

📚 Best Practices

1. Agent Selection

• Match agent capabilities to task requirements
• Consider agent load and availability
• Use specialized agents for domain-specific tasks
• Leverage swarm coordination for complex projects

2. Swarm Coordination

• Use hierarchical topology for large, structured projects
• Use mesh topology for collaborative, peer-level work
• Use adaptive topology for dynamic, changing requirements
• Monitor and adjust topology based on performance

3. Performance Optimization

• Batch similar tasks to reduce context switching
• Use memory coordination for knowledge sharing
• Monitor agent resource usage and optimize allocation
• Implement proper error handling and recovery

4. Quality Assurance

• Always include reviewer agents in production workflows
• Use multiple specialized agents for comprehensive coverage
• Implement consensus mechanisms for critical decisions
• Maintain audit trails for compliance and debugging

---

🚀 Future Roadmap

Planned Agent Types

• Quantum Computing Specialist: Quantum algorithm development
• IoT Coordinator: Internet of Things device management
• Blockchain Developer: Smart contract and DApp development
• AI Ethics Auditor: AI bias and ethics assessment
• Sustainability Analyzer: Carbon footprint and sustainability assessment

Enhanced Capabilities

• Multi-modal agent communication
• Advanced learning and adaptation
• Predictive task assignment
• Autonomous agent creation
• Cross-platform integration

---

📮 Support & Resources

Documentation

• Agent API Reference
• Swarm Coordination Guide
• Architecture Overview

Community

• GitHub Repository
• Discord Community
• Issue Tracker

Contact

• Email: support@claude-flow.ai
• Documentation: https://claude-flow.ai/docs
• Blog: https://blog.claude-flow.ai

---

🤖 65+ Specialized Agents • 🐝 Intelligent Swarm Coordination • 🚀 Enterprise-Ready

⬆ Back to Top