Skip to content

Introduction

Prodigy is an AI-powered workflow orchestration tool that enables development teams to automate complex tasks using Claude AI through structured YAML workflows.

What is Prodigy?

Prodigy combines the power of Claude AI with workflow orchestration to:

  • Automate repetitive development tasks - Code reviews, refactoring, testing
  • Process work in parallel - MapReduce-style parallel execution across git worktrees
  • Resume long-running operations - Checkpoint and resume capabilities for workflows that span hours or days
  • Handle failures gracefully - Dead Letter Queue (DLQ) for automated retry of failed items
  • Maintain quality - Built-in validation, error handling, and retry mechanisms
  • Track changes - Full git integration with automatic commits and merge workflows
  • Generate living documentation - Keep docs synchronized with code automatically
  • Compose complex workflows - Import, extend, and template reusable workflow components
graph TD
    User[User defines workflow.yml] --> Prodigy[Prodigy Orchestrator]
    Prodigy --> Parse[Parse & Validate YAML]
    Parse --> Worktree[Create Git Worktree]
    Worktree --> Execute[Execute Commands]

    Execute --> Shell[Shell Commands]
    Execute --> Claude[Claude AI Commands]
    Execute --> Control[Control Flow]

    Shell --> Commit[Git Commit]
    Claude --> Commit
    Control --> Commit

    Commit --> Checkpoint{Checkpoint?}
    Checkpoint -->|Save State| Resume[Resume Later]
    Checkpoint -->|Continue| Next{More Steps?}

    Next -->|Yes| Execute
    Next -->|No| Merge[Merge to Main Branch]
    Merge --> Complete[Complete]

    Resume -.->|prodigy resume| Execute

    style User fill:#e8f5e9
    style Prodigy fill:#e1f5ff
    style Claude fill:#fff3e0
    style Merge fill:#f3e5f5

Figure: Prodigy workflow execution showing orchestration, isolation, and checkpoint/resume capability.

Production-Ready Features

Prodigy includes enterprise features like checkpoints for resuming interrupted workflows, a Dead Letter Queue for automatic failure recovery, and git worktree isolation to keep your main repository clean during execution.

Quick Start

Your First Workflow

Start with a simple build-test cycle to see Prodigy in action. This example shows the simplified array syntax for straightforward workflows.

Create a simple workflow in workflow.yml:

# Source: workflows/complex-build-pipeline.yml (simplified)
# Simplified array syntax - use this for straightforward workflows
- shell: "cargo build"                    # (1)!
- shell: "cargo test"                     # (2)!
  on_failure:
    claude: "/fix-failing-tests"          # (3)!
- shell: "cargo clippy"                   # (4)!

1. Build the project
2. Run tests
3. If tests fail, Claude automatically attempts fixes
4. Run linting checks

Syntax Options

This example uses the simplified array syntax for simple workflows. For more complex workflows with parallel execution, MapReduce, or advanced features, use the full workflow structure with name, mode, and other configuration fields.

Run it:

prodigy run workflow.yml

Documentation Features

This book itself is maintained using Prodigy's automated documentation system! Learn how to set up automated, always-up-to-date documentation for your own project:

Key Concepts

  • Workflows: YAML files defining sequences of commands
  • Commands: Shell commands, Claude AI invocations, or control flow
  • Variables: Capture output from commands and use throughout workflow with ${VAR} syntax. Supports nested fields, defaults, and environment variables. See Environment Variables for details
  • Environment: Configuration with secrets management and profile-based values
  • MapReduce: Parallel processing across multiple git worktrees
  • Checkpoints: Save and resume workflow state for long-running operations
  • Validation: Workflow structure and syntax validation, implementation completeness checking with the validate command, and runtime validation. See Command Types for details
  • Observability: Event tracking, Claude execution logs, and comprehensive debugging tools

Start Simple, Scale Up

Begin with simple sequential workflows to learn the basics, then progress to MapReduce for parallel processing when you need to handle 10+ similar tasks concurrently.

Next Steps

Getting Started:

Advanced Features:

Operations: