Two things have been going super viral recently in how to use Claude Code in a structured way.

@agentic.james

19.9K followers · usually 1.3K views a post

Views

124.3K

Likes

7.0K

Comments

170

Shares

714

Selling · Course / Info

Agent Architects

https://www.skool.com/agent-architects ↗

Why it worked

The video provides high-value, technical 'how-to' content that solves a specific problem for developers, establishing the creator's authority and driving interest in his community.

Summary

The creator explains how to combine two specific frameworks, the 'Get Shit Done' framework and the 'Ralph Loop', to improve the output of Claude Code. He details how each system functions and how they can be used together to programmatically execute software development tasks.

Structure

1Introduce the two viral frameworks
2Explain the purpose of the Get Shit Done framework
3Explain the mechanism of the Ralph Loop
4Demonstrate how to combine both for better results

Product placement

The video discusses Claude Code, but the creator is positioning himself as an expert in AI agents, leading to his 'Agent Architects' community.

On-screen text

CLAUDE CODE GET SHIT DONE + RALPH LOOP

Original caption

How to use Claude Code Get Shit Done framework plus Ralph loop together to get the best outputs from Claude code. #claude #aiagents #vibecoding #aitools #ai

More from @agentic.james

Their other posts in the index, biggest breakout first.

Transcript

Two things have been going super viral recently in how to use Claude Code in a structured way. The first is the get shit done framework and the second is the Ralph loop. I'm going to tell you how to use these two things together to get the best outputs from Claude Code. First of all, we need to see that these two systems serve very different purposes in the actual execution workflow. The get shit done framework is really good for getting your idea and turning it into structured prompts or task documents to have Claude Code go through sequentially to execute your idea and turn it into an actual finalised software. Now the Ralph loop is an actual mechanism that uses Claude Code's internal hook structure to re-inject a certain prompt into Claude Code's context window over and over to get it to programmatically execute tasks until it reaches a certain completion state. So you can see how you can combine these two things together. We use the get shit done framework to get your idea out onto paper and out into structured task files for Claude Code to complete, and then actually use the Ralph loop to have Claude Code programmatically go through all of those task files and execute them one by one until all tests pass and your completion state is reached.