The Ralph Wiggum Technique
Join me as I perform a deep-dive into the agentic looping technique known as "Ralph Wiggum." In this series, we’re not just building a script, we’re building the surrounding system that makes an agent reliable enough to collaborate with. We’ll work through permissions, sandbox rules, hooks, and safer ways to run in more autonomous modes.
By the end, we’ll have something much bigger than a typical demo: a working approach for spec-driven agent development that can run locally, in the background, or on a remote machine over SSH. Let's go!
Series episodes
What Is Ralph?
5m 7sBuild a simple Ralph.sh loop and learn how lightweight memory files power agentic workflows to restart models and ship work faster.
Improving Our First Ralph Loop
4mImprove your first Ralph loop by replacing a shell script with a Node.js wrapper that parses stream-json into readable real-time output.
What Is A PRD?
6m 59sLearn what a PRD is and how to create an agent-ready PRD with clear context, constraints, and goals for better AI-built code.
Creating A PRD With Our PRD Skill
10m 18sLearn how to create a PRD that guides an AI agent through a real feature build, refine the plan, and test the results in your app.
Running Ralph In The Background
9m 8sLearn a simple tmux and screen workflow to keep long-running terminal tasks alive and your development setup resilient.
Permissions And Sandboxing
6m 46sLearn how to use allow/deny rules and sandboxing to reduce CLI noise, automate safely, and keep your model secure.
Configuring Hooks
7m 12sLearn how to use pre-tool-use hooks to secure MCP tools with smarter permission checks and model-driven review.
A Word On Auto Mode
3m 30sExplore Anthropic Auto mode, where it shines, where it feels shaky, and why a safer permission bypass approach still matters.
Bypassing Permissions Safely
10m 31sLearn how to use bypass permissions safely with sandboxes and Yolo box to reduce risk while exploring more isolated setups.
Turning Our Ralph Script Into a Flexible CLI
7m 40sTurn Ralph.js into a flexible CLI with Bun and OpenTUI. Build a tabbed terminal interface and streamline your PRD-driven workflow.
Agent Workspaces
9m 14sLearn how to isolate agent work with git worktree and copy-on-write so you can keep coding, test safely, and avoid main-branch conflicts.
Tracking PRDs With Linear
9m 46sLearn how to track PRDs in Linear, create parent and child issues, and automate pull requests for a faster dev workflow.
Human In The Loop
5m 44sLearn how to add a human-in-the-loop workflow that reviews AI architecture decisions and keeps your Laravel feature on track.
Offloading Loops To A Remote Machine
5m 37sLearn to run Ralph loops on a remote machine with SSH and tmux, keeping your workflow persistent, portable, and easy to control.
Experimenting With A Perpetual Loop
7m 17sSee what the perpetual loop cam experiment actually produced, how far Ralph got, and why agent orchestration matters next.
Beyond Ralph Orchestration
7m 59sBuild a flexible AI orchestration system with Linear, workflow.md files, Claude, and routing for a practical day-to-day agent workflow.
