How do I install the Ralph skill?

Run npx skillmds@latest add george-rd/ralph in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Ralph safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Ralph?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Ralph free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

george-rd (@george-rd) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Ralph by george-rd — Agent Skill

Name: Ralph
Author: george-rd

Ralph

Use for spec-driven development loops, porting code between languages, migrating codebases, retrospective analysis, or autonomous build loops. Trigger phrases: "ralph loop", "spec-driven", "build feature", "port", "clone to", "migrate to", "translate to", "rewrite in", "convert from X to Y", "autonomous loop", "implementation plan", "JTBD", "retrospective", "retro", "post-mortem", "what went wrong", "audit the project", "improvement todo".

george-rd c0febeb 1.6 KB Updated Sep 17, 2026

File contents

Ralph Wiggum Toolkit

Recipe-based autonomous development loops with spec-driven methodology.

Quick Start

Greenfield (new features)

Run /ralph init to scaffold the project
Write specs: /ralph spec
Plan: /ralph plan
Build: /ralph build

Port (codebase migration)

Run /ralph init --recipe port <source-lang> <target-lang> to scaffold
Run headless extraction loops for test and source specs
Plan: /ralph plan
Build: /ralph build

How It Works

Recipes define what gets scaffolded, which prompts exist, and how many phases
Loop infrastructure (loop.sh, stop-hook.sh, setup-loop.sh) is shared across all recipes
Manifest (ralph/manifest.json) tracks progress across iterations
Each iteration gets fresh context; continuity lives in files on disk

Available Recipes

greenfield: User-written specs → plan → build. For new features.
port: Extract behavioral specs → port to target language. For codebase migration.
retrospective: Analyze a completed Ralph project. Produces: codegap -> implementation gap -> plugin gap -> synthesis -> session explanation -> improvement TODO.

Full guide: /ralph help

Frequently asked questions