# Add to your Claude Code skills
git clone https://github.com/PageAI-Pro/ralph-loop
README.md
A Ralph Wiggum Loop implementation that works™
Ralph is a long-running AI agent loop. Ralph automates software development tasks by iteratively working through a task list until completion. This allows for long running agent loops, effectively enabling AI to code for days at a time.
This is an implementation that actually works, containing a hackable script so you can configure it to your env and favorite agentic AI CLI. It's set up by default to use Claude Code in a Docker sandbox, but supports many other agentic AI CLIs.
If you must start from a blank slate, which is not recommended, see Starting from scratch. You can also go for a more barebone start by running npx create-vite@latest src --template react-ts
1️⃣ Step 1: Install Ralph
Run this in your project's directory to install Ralph.
npx @pageai/ralph-loop
2️⃣ Step 2: Create a PRD + task list
Use the prd-creator skill to generate a PRD from your requirements.
Open up Claude Code or Cursor etc. and prompt it with your requirements. Like so:
Use the prd-creator skill to help me create a PRD and task list for the below requirements.
An app is already set up with React, Tailwind CSS and TypeScript.
Requirements:
- A SaaS product that helps users manage their finances.
- Target audience: Small business owners and freelancers.
- Core features:
- Track income and expenses.
- Create and send invoices.
- Track payments and receipts.
- Generate reports and insights.
- Connect to bank accounts and credit cards.
- Connect to accounting software.
- Connect to payment processors.
- Use the shadcn/ui library for components.
- Integrate with Stripe for payments.
- Use Supabase for database.
- You can find env variables in the .env.example file: SUPABASE_URL, SUPABASE_PUBLISHABLE_KEY, STRIPE_SECRET_KEY, etc. are available in the runtime.
// etc.
Check out the video for a more realistic example on how to write requirements.
mention libraries and frameworks you want to use
mention env variables, e.g. for DB, 3rd party API keys, etc. Store them in .env and add it to .gitignore
describe user flows and journeys
add relevant docs and UI references if applicable inside /docs and mention them in the requirements
be as descriptive as possible
it's fine to write in your own language
Then, follow the Skill's instructions and verify the PRD and then tasks.
It is highly recommended that you review individual task requirements before starting the loop. Review EACH TASK INDIVIDUALLY.
3️⃣ Step 3: Set up the agent inside Docker sandbox
Authenticate inside the Docker sandbox before running Ralph. Run:
docker sandbox run claude .
And follow the instructions to log in into Claude Code.
👉 Answer "Yes" to Bypass Permissions mode, that's the exact reason why you are using the Docker sandbox.
./ralph.sh -n 50 # Run Ralph Loop with 50 iterations
✍️ Note: the first iteration will be spent on ensuring the sandbox environment is set up correctly. Expect 5 minutes to complete.
Running the Ralph Loop with custom options
# Run the agent loop (default: 10 iterations)
./ralph.sh
# Run with custom iteration limit
./ralph.sh 5
./ralph.sh -n 5
./ralph.sh --max-iterations 5
# Run exactly one iteration
./ralph.sh --once
# Show help
./ralph.sh --help
NB: you might need to run chmod +x ralph.sh to make the script executable.
The default "mode" is "implementation". Depending on your use case, you might want to change .agent/PROMPT.md to a different mode, e.g. "refactor", "review", "test" etc.
⚠️ If you want to use a different language or testing framework, see below.
If you'd like to use a different language, testing framework etc. please adjust .agent/PROMPT.md to reflect your setup, server ports and startup commands etc.
👉 The loop is controlled by this prompt, which will be sent to the agent each iteration.
How It Works
Each iteration, Ralph will:
Find the highest-priority incomplete task from .agent/tasks.json
Work through the task steps defined in .agent/tasks/TASK-{ID}.json
Run tests, linting, and type checking
Complete task, take screenshot, update task status and commit changes
Repeat until all tasks pass or max iterations reached
How Is This Different from Other Ralphs?
This was kept hackable so you can make it your own.
The script follows the original concepts of the Ralph Wiggum Loop, working with fresh contexts and providing clear verifiable feedback.
It also works generically with any task set.
PRD generation - Creates a PRD and task list from requirements
Task lookup table generation - Creates a task lookup table from the PRD
Task breakdown + step generation - Breaks down each task into manageable steps
Iteration tracking - Shows progress through iterations with timing
Stream preview - Shows live output from the Agent
Step detection - Identifies current activity (Thinking, Implementing, Testing, etc.)
Screenshot capture - Captures a screenshot of the current screen
Notifications - Alerts when human input is needed
History logging - Saves clean output from each iteration
Timing - Shows timing metrics for each iteration and total time
Steering - Allows prioritizing critical work that needs to be done before the loop can continue
it allows you to dump unstructured requirements and have the agent create a PRD and task list for you.
it uses a task lookup table with individual detailed steps → more scalable as you get 100s of tasks done.
it's sandboxed and more secure
it shows progress and stats so you can keep an eye on what's been done
it instructs the agent to write and run automated tests and screenshots per task
it provides observability and traceability of the agent's work, showing a stream of output and capturing full historical logs per iteration
Steering the Agent
In some cases, you might notice the agent is having trouble, slowed down or struggling to overcome a blocker.
While the loop is running, you can edit the .agent/STEERING.md file to add critical work that needs to be done before the loop can continue.
The agent will check this file each iteration and if it finds any critical work, it will skip tasks and complete the critical work first.
Support
The ralph.sh script is designed to be hackable.
It is configured to use Claude Code in a Docker sandbox by default, but with a one-liner change you can change it to use any other agentic AI CLI.
Check the ralph.sh script around # This is the main command loop. for the main command loop.
NB: skills are supported by all major agentic AI CLIs via symlinks.
Promise Tags
Ralph uses semantic tags to communicate status:
<promise>COMPLETE</promise> - All tasks finished successfully
<promise>BLOCKED:reason</promise> - Agent needs human help
<promise>DECIDE:question</promise> - Agent needs a decision
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
