ralph

Resolve the highest-priority issue from state.json using a TDD workflow.

38|6|Updated Sep 13, 2012
One-click install
npx skills add https://github.com/pixelastic/oroshi --skill ralph-pixelastic
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/pixelastic/oroshi/tree/main/tools/ai/claude/config/skills/ralph
Command: npx skills add https://github.com/pixelastic/oroshi --skill ralph-pixelastic

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill automates resolving the highest-priority issue from a given state.json directory using a TDD workflow, updates documentation, runs a review in a subagent, and pauses for user confirmation before committing.

Core Features & Use Cases

  • Orchestrates selecting the next issue from state.json and loading the GUIDANCE.md context for the session.
  • Runs a TDD cycle: sets up context, writes tests, and verifies fixes against the issue specification.
  • Initiates the review process and applies actionable feedback, then halts to await user commit.
  • Keeps project docs in sync by updating guidance and related materials as needed.

Quick Start

Invoke Ralph with a directory containing state.json and GUIDANCE.md to process the next issue and stop for user review.

Frequently Asked Questions about ralph

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate resolving the highest-priority issue from state.json using a TDD workflow?

You can automate issue resolution by pointing the Skill at a directory containing state.json and GUIDANCE.md. It selects the top-priority issue, loads guidance, runs a red/green TDD cycle, applies fixes, and pauses for user commit.

What is the process for loading issue specifications and guidance before writing tests?

The process involves reading GUIDANCE.md to load project context and extracting the issue specification from state.json. This context is loaded automatically before orchestrating the test setup and implementation phases.

How do I run a code review and apply feedback automatically before committing?

The workflow initiates a review process in a subagent after the TDD cycle completes. It applies actionable feedback from the review and then halts to await your confirmation before performing a non-automatic commit.

Can I use this TDD workflow if my project tracks issues in state.json without a GUIDANCE.md file?

No, you need both state.json for issue tracking and GUIDANCE.md for loading session context. The workflow relies on these specific files to select the next issue and apply the appropriate project guidance during the TDD cycle.

How does this approach keep project documentation in sync with code changes?

After completing the implementation and review phases, the workflow automatically updates guidance materials and related project documentation to ensure the docs remain in sync with the applied fixes before pausing for commit.

When should I expect the automated issue workflow to pause for manual intervention?

The workflow pauses for manual intervention right before the final commit step. It performs a non-automatic commit to ensure you can review the completed TDD cycle, applied fixes, and documentation updates before changes are finalized.