Can I review changes before they are finalized?

Yes, by using the --interactive flag, the skill will pause after each iteration to show a summary and wait for your confirmation before proceeding to the next subtask.

Ralph Build is a skill for Claude Code that executes an autonomous loop to process a list of subtasks defined in a subtasks.json file, following a standardized development workflow.

How does Ralph Build handle task failures?

The skill tracks retry counts per subtask and will stop after a specified number of failures (defaulting to 3) to prevent infinite loops and allow for manual intervention.

Where does the skill look for subtasks?

If no path is provided, Ralph Build automatically searches for subtasks.json files within the docs/planning/milestones/ directory structure of your project.

Is there a way to test the build loop without making changes?

Yes, the --print or -p option serves as a dry-run mode, outputting the prompts that would be sent to Claude without actually executing any iterations.

Ralph Build Loop

Name: Ralph Build Loop
Author: otrebu

byotrebu

•

Productividad y Flujo de Trabajo

Automates the execution of subtask queues through an autonomous build loop for structured software development.

Acerca de

Ralph Build is a specialized Claude Code skill designed to process subtasks autonomously from a structured JSON queue. It streamlines the development process by iterating through planned milestones, offering features like interactive review pauses, pre-build validation, and configurable retry limits. This skill is ideal for developers who want to delegate the mechanical execution of a task list to Claude while maintaining control over the progression and quality of each iteration through defined workflows.

Características Principales

Dry-run support to preview prompts and logic before execution
Autonomous subtask processing from JSON-based task queues
Pre-build validation to ensure environment and code readiness
Configurable retry logic with customizable maximum iteration limits
3 GitHub stars
Interactive mode for manual review between task iterations

Casos de Uso

Executing complex development milestones from a structured planning document
Managing autonomous coding loops with human-in-the-loop oversight
Automating repetitive implementation tasks across multiple project files

Acerca de

Características Principales

Dry-run support to preview prompts and logic before execution
Autonomous subtask processing from JSON-based task queues
Pre-build validation to ensure environment and code readiness
Configurable retry logic with customizable maximum iteration limits
3 GitHub stars
Interactive mode for manual review between task iterations

Casos de Uso

Executing complex development milestones from a structured planning document
Managing autonomous coding loops with human-in-the-loop oversight
Automating repetitive implementation tasks across multiple project files