Ralph Wiggum：基于规范驱动开发的自主AI编程工具，解决上下文窗口溢出

ralph-wiggum by fstandhartinger/ralph-wiggum

590 周安装量

201 GitHub Stars

GitHub

安装命令

npx skills add https://github.com/fstandhartinger/ralph-wiggum --skill ralph-wiggum

AI/机器学习开发自动化

🇨🇳中文介绍

Ralph Wiggum

基于规范驱动开发的自主式 AI 编程

什么是 Ralph Wiggum？

Ralph Wiggum 将 Geoffrey Huntley 的迭代式 bash 循环 与 规范驱动开发 相结合，实现完全自主的 AI 辅助软件开发。

核心洞见：每次迭代都有全新的上下文。每个循环都会启动一个具有干净上下文窗口的新代理进程，从而防止上下文溢出和性能下降。

何时使用此技能

在以下情况下使用 Ralph Wiggum：

你有多个规范/功能需要实现
你希望 AI 能自主处理任务
你需要可验证地、一致地完成验收标准
你希望避免长会话中的上下文窗口问题

工作原理

┌─────────────────────────────────────────────────────────────┐
│                     RALPH LOOP                              │
├─────────────────────────────────────────────────────────────┤
│  Loop 1: Pick spec A → Implement → Test → Commit → DONE    │
│  Loop 2: Pick spec B → Implement → Test → Commit → DONE    │
│  Loop 3: Pick spec C → Implement → Test → Commit → DONE    │
│  ...                                                        │
│                                                             │
│  Each iteration = Fresh context window                      │
│  Shared state = Files on disk (specs, plan, history)        │
└─────────────────────────────────────────────────────────────┘

广告位招租

在这里展示您的产品或服务

触达数万 AI 开发者，精准高效

联系我们

快速安装（通过技能安装器）

# Using Vercel's add-skill
npx add-skill fstandhartinger/ralph-wiggum

# Using OpenSkills
openskills install fstandhartinger/ralph-wiggum

完整设置（推荐）

要进行包含章程和项目访谈的完整 Ralph Wiggum 设置：

# Tell your AI agent:
"Set up Ralph Wiggum using https://github.com/fstandhartinger/ralph-wiggum"

代理将引导你完成一个 轻量级、愉快的设置过程：

快速设置（约 1 分钟）—— 创建目录，下载脚本
项目访谈 —— 专注于你的 愿景和目标（而非技术细节）
章程 —— 创建一个指导所有会话的文档
后续步骤 —— 提供关于创建规范和启动 Ralph 的清晰指导

对于现有项目，代理会自动检测你的技术栈。访谈优先关注 你正在构建什么以及为什么。

1. 每次循环都有全新上下文

Ralph 循环的每次迭代都会启动一个新的 AI 代理进程。这意味着：

没有上下文窗口溢出
不会随时间推移而性能下降
每个任务都是全新的开始

2. 基于磁盘的共享状态

状态通过文件在循环之间持久化：

specs/ —— 包含验收标准的特性规范
ralph_history.txt —— 记录突破、阻碍和学习的日志
IMPLEMENTATION_PLAN.md —— 可选的详细任务分解

代理仅在以下情况下输出 <promise>DONE</promise>：

所有验收标准均已验证
测试通过
更改已提交并推送

bash 循环会检查这个短语。如果未找到，则重试。

4. 通过测试实现反压

测试、代码检查和构建充当了护栏。代理必须在输出完成信号之前修复问题。

成功的关键： 每个规范都需要 清晰、可测试的验收标准。这能告诉 Ralph 任务何时真正“完成”。

# Feature: User Authentication

## Requirements
- OAuth login with Google
- Session management
- Logout functionality

## Acceptance Criteria
- [ ] User can log in with Google
- [ ] Session persists across page reloads
- [ ] User can log out
- [ ] Tests pass

**Output when complete:** `<promise>DONE</promise>`

好的标准： “用户可以使用 Google 登录且会话保持” 坏的标准： “认证功能正常工作”

你的验收标准越具体，Ralph 的表现就越好。

# Start building (Claude Code)
./scripts/ralph-loop.sh

# With max iterations
./scripts/ralph-loop.sh 20

# Using Codex CLI
./scripts/ralph-loop-codex.sh

日志记录（捕获所有输出）

每次循环运行都会将 所有输出 写入 logs/ 目录下的日志文件：

会话日志： logs/ralph_*_session_YYYYMMDD_HHMMSS.log（整个运行过程，包括 CLI 输出）
迭代日志： logs/ralph_*_iter_N_YYYYMMDD_HHMMSS.log（每次迭代的 CLI 输出）
Codex 最后消息： logs/ralph_codex_output_iter_N_*.txt

模式	用途	命令
build（默认）	选择规范、实现、测试、提交	`./scripts/ralph-loop.sh`
plan（可选）	创建详细的任务分解	`./scripts/ralph-loop.sh plan`

信任 AI 能够自我识别、自我纠正和自我改进。观察模式并调整提示。

为了让 Ralph 有效工作，请启用完全自主权限：

Claude Code: --dangerously-skip-permissions
Codex: --dangerously-bypass-approvals-and-sandbox

⚠️ 风险自负。 仅在沙盒环境中使用。

🇺🇸English

Ralph Wiggum

Autonomous AI coding with spec-driven development

What is Ralph Wiggum?

Ralph Wiggum combines Geoffrey Huntley's iterative bash loop with spec-driven development for fully autonomous AI-assisted software development.

The key insight: Fresh context each iteration. Each loop starts a new agent process with a clean context window, preventing context overflow and degradation.

When to Use This Skill

Use Ralph Wiggum when:

You have multiple specifications/features to implement
You want the AI to work autonomously through tasks
You need consistent, verifiable completion of acceptance criteria
You want to avoid context window problems in long sessions

How It Works

┌─────────────────────────────────────────────────────────────┐
│                     RALPH LOOP                              │
├─────────────────────────────────────────────────────────────┤
│  Loop 1: Pick spec A → Implement → Test → Commit → DONE    │
│  Loop 2: Pick spec B → Implement → Test → Commit → DONE    │
│  Loop 3: Pick spec C → Implement → Test → Commit → DONE    │
│  ...                                                        │
│                                                             │
│  Each iteration = Fresh context window                      │
│  Shared state = Files on disk (specs, plan, history)        │
└─────────────────────────────────────────────────────────────┘

Installation

Quick Install (via Skill Installers)

# Using Vercel's add-skill
npx add-skill fstandhartinger/ralph-wiggum

# Using OpenSkills
openskills install fstandhartinger/ralph-wiggum

Full Setup (Recommended)

For full Ralph Wiggum setup with constitution and interview:

# Tell your AI agent:
"Set up Ralph Wiggum using https://github.com/fstandhartinger/ralph-wiggum"

The agent will guide you through a lightweight, pleasant setup :

Quick Setup (~1 min) — Create directories, download scripts
Project Interview — Focus on your vision and goals (not tech details)
Constitution — Create a guiding document for all sessions
Next Steps — Clear guidance on creating specs and starting Ralph

For existing projects, the agent detects your tech stack automatically. The interview prioritizes understanding what you're building and why.

Core Concepts

1. Fresh Context Each Loop

Each iteration of the Ralph loop starts a new AI agent process. This means:

No context window overflow
No degradation over time
Clean slate for each task

2. Shared State on Disk

State persists between loops via files:

specs/ — Feature specifications with acceptance criteria
ralph_history.txt — Log of breakthroughs, blockers, learnings
IMPLEMENTATION_PLAN.md — Optional detailed task breakdown

3. Completion Signal

The agent outputs <promise>DONE</promise> ONLY when:

All acceptance criteria are verified
Tests pass
Changes are committed and pushed

The bash loop checks for this phrase. If not found, it retries.

4. Backpressure via Tests

Tests, lints, and builds act as guardrails. The agent must fix issues before outputting the completion signal.

Usage

Creating Specifications

The key to success: Each spec needs clear, testable acceptance criteria. This is what tells Ralph when a task is truly "done."

# Feature: User Authentication

## Requirements
- OAuth login with Google
- Session management
- Logout functionality

## Acceptance Criteria
- [ ] User can log in with Google
- [ ] Session persists across page reloads
- [ ] User can log out
- [ ] Tests pass

**Output when complete:** `<promise>DONE</promise>`

Good criteria: "User can log in with Google and session persists" Bad criteria: "Auth works correctly"

The more specific your acceptance criteria, the better Ralph performs.

Running the Loop

# Start building (Claude Code)
./scripts/ralph-loop.sh

# With max iterations
./scripts/ralph-loop.sh 20

# Using Codex CLI
./scripts/ralph-loop-codex.sh

Logging (All Output Captured)

Every loop run writes all output to log files in logs/:

Session log: logs/ralph_*_session_YYYYMMDD_HHMMSS.log (entire run, including CLI output)
Iteration logs: logs/ralph_*_iter_N_YYYYMMDD_HHMMSS.log (per-iteration CLI output)
Codex last message: logs/ralph_codex_output_iter_N_*.txt

Two Modes

Mode	Purpose	Command
build (default)	Pick spec, implement, test, commit	`./scripts/ralph-loop.sh`
plan (optional)	Create detailed task breakdown	`./scripts/ralph-loop.sh plan`

Key Principles

Let Ralph Ralph

Trust the AI to self-identify, self-correct, and self-improve. Observe patterns and adjust prompts.

YOLO Mode

For Ralph to work effectively, enable full autonomy:

Claude Code: --dangerously-skip-permissions
Codex: --dangerously-bypass-approvals-and-sandbox

⚠️ Use at your own risk. Only in sandboxed environments.

Ralph Wiggum：基于规范驱动开发的自主AI编程工具，解决上下文窗口溢出

🇨🇳中文介绍

Ralph Wiggum

什么是 Ralph Wiggum？

何时使用此技能

工作原理

相关 Skills

安装

快速安装（通过技能安装器）

完整设置（推荐）

核心概念

1. 每次循环都有全新上下文

2. 基于磁盘的共享状态

3. 完成信号

4. 通过测试实现反压

使用方法

创建规范

运行循环

日志记录（捕获所有输出）

两种模式

关键原则

让 Ralph 做 Ralph

YOLO 模式

链接

🇺🇸English

Ralph Wiggum

What is Ralph Wiggum?

When to Use This Skill

How It Works

Installation

Quick Install (via Skill Installers)

Full Setup (Recommended)

Core Concepts

1. Fresh Context Each Loop

2. Shared State on Disk

3. Completion Signal

4. Backpressure via Tests

Usage

Creating Specifications

Running the Loop

Logging (All Output Captured)

Two Modes

Key Principles

Let Ralph Ralph

YOLO Mode

Links

最新 Skills