This skill implements the 'Ralph' autonomous workflow, a system designed to allow Claude to work through a queue of implementation stories independently. By establishing a rigid structure consisting of a PRD, instruction prompt, progress log, and execution script, it enables Claude to implement, validate, and commit code overnight or in the background. It is specifically optimized for well-defined, small-scale tasks with clear acceptance criteria and automated test suites, providing a persistent memory mechanism through codebase pattern logging to ensure high-quality, consistent output without constant human intervention.
Key Features
01Creates a prompt.md with explicit validation commands and workflow instructions.
02Enforces task-sizing rules to ensure stories fit within AI context windows.
03Generates a structured prd.json for objective task tracking and acceptance criteria.
04Sets up a ralph.sh bash script to manage the autonomous execution loop.
05Initializes a progress.txt log to capture codebase patterns and persistent memory.
060 GitHub stars
