Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
Bradley Golden avatar

Ralph

  • Updated January 14, 2026
  • bradleygolden/ralph-claude-plugin

ralph is a Claude Code skill in the AI & Agent Building category. Ralph Wiggum loop - continuous Claude execution with repomirror visualization

Key points

  • ralph
  • AI & Agent Building
  • AI-coding skill

Ralph by the numbers

  • Data as of Jul 7, 2026 (Skillselion catalog sync)
/plugin marketplace add bradleygolden/ralph-claude-plugin
/plugin install ralph@ralph-claude-plugin

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Last updatedJanuary 14, 2026
Repositorybradleygolden/ralph-claude-plugin

What it does

Ralph Wiggum loop - continuous Claude execution with repomirror visualization

README.md

Ralph Claude Plugin

A Claude Code plugin that runs the "Ralph Wiggum loop" - continuous Claude execution with repomirror visualization.

What it does

The Ralph loop continuously:

  1. Executes Claude with your prompt in piped mode
  2. Pipes output to repomirror visualize for visualization
  3. Waits 10 seconds
  4. Repeats until you stop it with Ctrl+C

Installation

From GitHub

# Step 1: Add the marketplace
/plugin marketplace add bradleygolden/ralph-claude-plugin

# Step 2: Install the plugin
/plugin install ralph@ralph-claude-plugin --scope user

Local Development

claude --plugin-dir ./plugins/ralph

Usage

Default (sandbox mode)

Runs with sandbox enabled and auto-allow for bash commands:

/ralph:ralph "Review the codebase and suggest improvements"
/ralph:ralph @PROMPT.md

Unsafe mode (skip permissions)

Use --unsafe to bypass sandbox and use --dangerously-skip-permissions:

/ralph:ralph --unsafe "your prompt"
/ralph:ralph --unsafe @PROMPT.md

Specify a model

Use --model to specify which Claude model to use:

/ralph:ralph --model opus "your prompt"
/ralph:ralph --model haiku "quick task"
/ralph:ralph --model sonnet @PROMPT.md

Valid model values: sonnet, opus, haiku, or full model IDs.

Combine flags

Flags can be combined in any order:

/ralph:ralph --unsafe --model opus "your prompt"
/ralph:ralph --model haiku --unsafe @PROMPT.md

How to stop

Press Ctrl+C to stop the loop.

Requirements

  • repomirror - for visualization (npx repomirror visualize)
  • Claude CLI

License

MIT

Related skills

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.