Skip to content

Examples

Real-world workflow patterns you can copy and adapt.

Included Examples

Example Description Key Features
Simple Sequential Basic multi-step pipeline Dependencies, templates
Code Review Pipeline Multi-expert review Fan-out/fan-in, parallel
Security Scan Security-focused analysis Tool restrictions, severity
Decision Helper Interactive decision-making Multiple perspectives

Running Examples

All examples are in the examples/ directory:

# List available examples
ls examples/

# Run an example
goflow run --workflow examples/simple-sequential.yaml \
  --inputs files='pkg/**/*.go' \
  --mock \
  --verbose

Example Structure

Each example follows this pattern:

examples/
└── example-name/
    ├── example-name.yaml     # Workflow file
    ├── README.md             # Documentation
    └── agents/               # Supporting agents
        ├── agent-a.agent.md
        └── agent-b.agent.md

Quick Pattern Reference

Sequential Pipeline

steps:
  - id: step-a
    prompt: "First task"
  - id: step-b
    prompt: "Second task using {{steps.step-a.output}}"
    depends_on: [step-a]

Parallel Processing

steps:
  - id: setup
    prompt: "Prepare"

  - id: task-1
    depends_on: [setup]
  - id: task-2
    depends_on: [setup]  # Runs parallel with task-1
  - id: task-3
    depends_on: [setup]  # Runs parallel with task-1, task-2

  - id: combine
    depends_on: [task-1, task-2, task-3]  # Waits for all

Conditional Branching

steps:
  - id: classify
    prompt: "Is this CRITICAL or NORMAL?"

  - id: urgent-path
    depends_on: [classify]
    condition:
      step: classify
      contains: "CRITICAL"

  - id: normal-path
    depends_on: [classify]
    condition:
      step: classify
      contains: "NORMAL"

Multi-Agent Collaboration

agents:
  expert-a:
    file: "./agents/security.agent.md"
  expert-b:
    file: "./agents/performance.agent.md"
  aggregator:
    inline:
      prompt: "You combine expert opinions."

steps:
  - id: review-a
    agent: expert-a
    prompt: "Analyze from your perspective"

  - id: review-b
    agent: expert-b
    prompt: "Analyze from your perspective"

  - id: combine
    agent: aggregator
    prompt: |
      Expert A says: {{steps.review-a.output}}
      Expert B says: {{steps.review-b.output}}

      Synthesize the findings.
    depends_on: [review-a, review-b]

Browse Examples