Claude Code subagent imported from bishop-l-walker-usps/icda-prototype (.claude/agents/README.md). Copyright stays with the author.

🤖 Specialized AI Agents

Domain-expert agents for focused technical assistance

📋 Available Agents

Each agent is a specialized expert with deep knowledge in their domain:

Technology Domain Agents

Data Platform & MCP Agents

Code Refactoring & Quality Agents

🎯 How to Use Agents

Method 1: Slash Commands

# Technology Domain Agents
/agent-kafka
/agent-aws
/agent-springboot
/agent-springcloud
/agent-docker

# Data Platform & MCP Agents
/agent-genie         # Databricks Genie API
/agent-mcp           # MCP server architecture
/agent-sql-security  # SQL security & governance

# Code Refactoring & Quality Agents
/agent-preserve      # Functionality preservation
/agent-sentinel      # Code quality analysis
/agent-depgraph      # Dependency graph
/agent-debt          # Technical debt
/agent-deadcode      # Dead code hunting
/agent-redundancy    # Redundancy elimination
/agent-consolidate   # Code consolidation
/agent-document      # Documentation (Ultrathink)

Method 2: Direct Context Load

Load the agent context file directly:

# Example: Loading Kafka agent
cat .claude/agents/KAFKA_AGENT.md

Method 3: Combined Context

Load multiple agents for cross-domain work:

# Example: Spring Boot + Kafka + Docker
cat .claude/agents/SPRINGBOOT_AGENT.md \
    .claude/agents/KAFKA_AGENT.md \
    .claude/agents/DOCKER_AGENT.md

🧠 Agent Capabilities

Kafka Agent

• Topic design and partitioning strategies
• Producer/Consumer configuration
• Kafka Streams and ksqlDB
• Schema Registry and Avro
• Performance tuning and monitoring
• Common issues and troubleshooting

AWS Agent

• Service selection and architecture
• IAM roles and security best practices
• Serverless (Lambda, API Gateway, EventBridge)
• Storage (S3, EFS, EBS)
• Database services (RDS, DynamoDB, Aurora)
• Networking (VPC, Load Balancers, CloudFront)
• Cost optimization strategies

Spring Boot Agent

• RESTful API design patterns
• Spring Data JPA and repositories
• Spring Security and authentication
• Configuration management (@ConfigurationProperties)
• Exception handling and validation
• Testing (MockMvc, TestRestTemplate)
• Actuator and monitoring

Spring Cloud Agent

• Microservices architecture patterns
• Service discovery (Eureka, Consul)
• Config Server centralization
• API Gateway (Spring Cloud Gateway)
• Circuit breakers (Resilience4j)
• Distributed tracing (Sleuth, Zipkin)
• Load balancing and fault tolerance

Docker Agent

• Dockerfile best practices
• Multi-stage builds for optimization
• docker-compose orchestration
• Networking and volumes
• Security hardening
• CI/CD integration
• Kubernetes basics

Databricks Genie Agent

• Genie Conversation API integration
• Multi-cloud authentication (AWS, Azure, GCP)
• Async polling and rate limiting
• MCP server creation for Genie
• LangChain integration
• Query result caching

MCP Architect Agent

• MCP protocol fundamentals
• Tool and resource design
• Multi-agent pipeline orchestration
• Quality gates and validation
• Docker containerization
• Client integration (Claude Desktop, VS Code)

SQL Security Agent

• SQL injection prevention
• PII detection and masking
• Unity Catalog row/column security
• OAuth 2.0 token management
• Query guardrails and cost controls
• Compliance frameworks (SOC 2, GDPR, HIPAA)

📚 Agent Context Structure

Each agent includes:

1. Core Concepts - Fundamental knowledge
2. Common Patterns - Best practices and design patterns
3. Code Examples - Ready-to-use snippets
4. Configuration - Common config patterns
5. Troubleshooting - Common issues and solutions
6. Performance Tips - Optimization strategies
7. Security - Security best practices
8. Testing - Testing strategies and examples
9. Resources - Documentation links and references

🔄 Agent Workflow

1. Identify Domain - Which technology are you working with?
2. Activate Agent - Use slash command or load context
3. Ask Questions - Agent responds with specialized knowledge
4. Implement Solution - Agent provides code and guidance
5. Switch Agents - Activate different agent as needed

🎓 Multi-Agent Scenarios

Scenario 1: Building a Microservice

Agents needed: Spring Boot + Spring Cloud + Docker

/agent-springboot
# Discuss: Create REST API with Spring Boot
/agent-springcloud
# Discuss: Add service discovery and config server
/agent-docker
# Discuss: Containerize the application

Scenario 2: Event-Driven Architecture

Agents needed: Kafka + Spring Boot + AWS

/agent-kafka
# Discuss: Design event topics and schemas
/agent-springboot
# Discuss: Implement Kafka producer/consumer
/agent-aws
# Discuss: Deploy to AWS with MSK (Managed Kafka)

Scenario 3: Cloud Deployment

Agents needed: Docker + AWS

/agent-docker
# Discuss: Build optimized container image
/agent-aws
# Discuss: Deploy to ECS/EKS with proper networking

Scenario 4: Databricks Genie MCP Server

Agents needed: Genie + MCP + SQL Security

/agent-genie
# Discuss: Understand Genie API and polling patterns
/agent-mcp
# Discuss: Design MCP server with tools
/agent-sql-security
# Discuss: Add query validation and PII protection

📊 Agent Selection Guide

Use this flowchart to pick the right agent:

Working with...
├── Message queues/events? → Kafka Agent
├── Cloud infrastructure? → AWS Agent
├── Building APIs/services? → Spring Boot Agent
├── Microservices coordination? → Spring Cloud Agent
├── Containers/deployment? → Docker Agent
├── Databricks NL-to-SQL? → Genie Agent
├── MCP server development? → MCP Architect Agent
└── SQL security/governance? → SQL Security Agent

🚀 Pro Tips

1. Start Broad, Get Specific - Begin with general architecture questions, then dive into specifics
2. One Agent at a Time - Focus on one domain, then switch contexts
3. Combine for Integration - Load multiple agents when working on integration points
4. Reference Agent Docs - Each agent has comprehensive examples and patterns
5. Ask for Code - Agents provide production-ready code snippets

📝 Creating Custom Agents

You can create your own specialized agents:

1. Copy the agent template: AGENT_TEMPLATE.md
2. Fill in domain-specific knowledge
3. Add to this README
4. Create a slash command in .claude/commands/

🔗 Related Documentation

• Main Context: .claude/CLAUDE.md - General development standards
• Project Setup: .claude/INITIAL.md - Project specifications
• Architecture: PLANNING.md - System architecture
• Task Tracking: TASK.md - Current tasks

🧹 Code Cleanup Pipeline

The refactoring agents work together in a pipeline for safe code cleanup:

RECOMMENDED WORKFLOW:

1. /agent-preserve   → Create functionality baseline BEFORE changes
2. /agent-deadcode   → Find all dead/unused code
3. /agent-redundancy → Find all duplicate code
4. /agent-consolidate → Execute cleanup safely
5. /agent-preserve   → Verify no functionality lost
6. /agent-document   → Update documentation

TARGET: 40% codebase reduction while maintaining ALL functionality

Code Cleanup Agent Capabilities

Status: ✅ Active Total Agents: 16 (5 technology + 3 data platform/MCP + 8 refactoring) Last Updated: 2025-12-17