#374374th of 389 in Skills, plugins and prompts
open-ralph-wiggum
Loop runner that re-prompts a coding agent until the task is done
- Stars
- 1.9k
- License
- MIT
- Last commit
- Jun 2026
- Last release
- May 2026
- Language
- TypeScript
1,525 on 21 Apr 20261,906 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Open Ralph Wiggum is a Bun and TypeScript CLI (`ralph`) that implements the Ralph Wiggum loop. It sends the same prompt to a coding agent each iteration until the agent outputs a completion promise or hits a max-iteration limit. It wraps Claude Code, Codex, Copilot CLI, Cursor Agent, Qwen Code and OpenCode, and adds a tasks mode, status monitoring and mid-loop context injection.
Who it is for: Developers running unattended coding-agent loops from the terminal
Strengths
- Switches between six agent CLIs with one --agent flag
- Tasks mode tracks progress in .ralph/ralph-tasks.md and survives restarts
- --status and --add-context let you monitor and steer a running loop from another terminal
- Custom prompt templates with variables for iteration, tasks and injected context
Weaknesses
- Requires Bun and a separately installed agent CLI
- Auto-approves all tool permissions by default (--allow-all is on)
- Max iterations default to unlimited, so a run can use tokens without a cap
- Auto-commits after iterations unless --no-commit is passed
What it needs
- no GPU
- Needs Bun, At least one agent CLI: Claude Code, Codex, Copilot CLI, Cursor Agent, Qwen Code or OpenCode
- Models: Claude Code, OpenAI Codex, Copilot CLI, Cursor Agent, Qwen Code
Quick answers
Is open-ralph-wiggum open source?
Yes. Its code is on GitHub under the MIT license.
Also in Skills, plugins and prompts
See all 389| Rank | Project | Score |
|---|---|---|
| 1 | claude-memPersistent cross-session memory plugin for Claude Code and other coding agents | 97 out of 100 |
| 2 | impeccableDesign skill with 24 slash commands for AI coding agents' frontend output | 97 out of 100 |
| 3 | gstackMarkdown slash-command skills that give Claude Code planning, review, QA and ship roles | 96 out of 100 |
| #4 | rufloAgent harness adding swarms, memory and hooks to Claude Code and Codex | 96 out of 100 |
| #5 | cavemanShortens coding-agent replies and compresses tool output before the model reads it | 95 out of 100 |