ContextQMD
Back to Ruflo

Workflow Examples

v2/examples/02-workflows/README.md

3.6.303.2 KB
Original Source

Workflow Examples

This directory contains multi-agent workflow definitions demonstrating various coordination patterns and use cases.

Directory Structure

02-workflows/
├── simple/         # Basic workflows for learning
├── parallel/       # Parallel execution patterns
├── sequential/     # Step-by-step workflows
├── complex/        # Advanced multi-agent systems
├── specialized/    # Domain-specific workflows
├── claude-workflow.json      # Original development workflow
└── research-workflow.json    # Original research pipeline

Workflow Categories

Simple Workflows (simple/)

  • hello-world-workflow.json: Single-agent starter workflow
    bash
    cd examples
../claude-flow swarm create "Build hello world app" --output ./output/hello-world

Parallel Workflows (parallel/)

  • data-processing-workflow.json: Process multiple data sources simultaneously
    bash
    ../claude-flow swarm create "Process CSV, JSON, and XML data in parallel" --agents 4 --output ./output/data-processing

Sequential Workflows (sequential/)

  • blog-platform-workflow.json: Step-by-step blog platform development
    bash
    ../claude-flow swarm create "Build complete blog platform with authentication" --strategy development --output ./output/blog

Complex Workflows (complex/)

  • microservices-workflow.json: Complete microservices architecture with 8 agents
    • System design → Service development → Frontend → DevOps → Testing
    • Smart execution with checkpoints and rollback
    bash
    ../claude-flow swarm create "Build microservices e-commerce platform" --agents 8 --output ./output/microservices

Specialized Workflows (specialized/)

  • machine-learning-workflow.json: End-to-end ML pipeline
    • Data prep → Feature engineering → Model research → Training → Deployment
    • Experiment tracking and model versioning
    bash
    ../claude-flow swarm create "Build machine learning pipeline for customer churn prediction" --strategy analysis --output ./output/ml-pipeline

Legacy Workflows

claude-workflow.json

Original multi-agent development workflow example

research-workflow.json

Original AI research pipeline example

Workflow Structure

Agents Section

json
"agents": [
  {
    "id": "agent-name",
    "type": "researcher|developer|tester",
    "capabilities": ["research", "code-generation", "testing"],
    "configuration": { ... }
  }
]

Tasks Section

json
"tasks": [
  {
    "id": "task-id",
    "agentId": "agent-name",
    "type": "research|coding|analysis",
    "dependencies": ["previous-task-id"],
    "parallel": true|false
  }
]

Running Workflows

bash
# Execute a workflow
npx claude-flow orchestrate ./claude-workflow.json

# With monitoring
npx claude-flow orchestrate ./research-workflow.json --monitor

# In background
npx claude-flow orchestrate ./claude-workflow.json --background

Creating Custom Workflows

  1. Define your agents with specific capabilities
  2. Create tasks assigned to appropriate agents
  3. Set up dependencies between tasks
  4. Configure parallel execution where possible
  5. Add quality thresholds and validation steps