This Claude Code Plugin Transforms Vibe Coders Into Senior Engineers

Video Description

Join the AI MBA to learn personally from me how to build with Claude Code ⤵️ https://www.skool.com/the-ai-mba Prefer reading tutorials to watching them? Check out my written guides ⤵️ https://ai-mba.io/tutorials Summary ⤵️ Every vibe coder has experienced a Claude Code project quickly turning into spaghetti code. This video introduces compound engineering - a 4-step process (Plan, Work, Review, Compound) that transforms vibe coders into senior engineers. I walk through Every's compound engineering plugin that automates this workflow, and demonstrate it live on a real paying client project. Links 🚀 Compound Engineering Plugin: https://github.com/EveryInc/compound-engineering-plugin Every (creators of compound engineering): https://every.to Follow me on other platforms 😈 🕊️ Twitter/X: https://x.com/aniketapanjwani 💼 LinkedIn: https://linkedin.com/in/aniket-a-panjwani 🤙 Work with me: https://contentquant.io Why watch? If this is your first view—hi, I'm Aniket! TLDR: I've worked for six years as an MLOps engineer at Zelle building financial AI models, automations and data pipelines saving consumers from tens of millions of dollars of fraud. Now, I use my professional and academic expertise (PhD Economics from Northwestern University) to help businesses automate tedious but critical processes to drive revenue and innovation. Hopefully I can help you improve your business, and in doing so, the rest of your life 🙏 Like, subscribe, and leave me a comment if you have a specific request! Thanks. 0:00 Every vibe coder has experienced spaghetti code 0:22 What is compound engineering? 0:42 The four-step process: Plan, Work, Review, Compound 1:17 Step 2: Work - let the agent execute your plan 2:01 Step 3: Review - specialized subagents check your code 3:56 Step 4: Compound - document learnings for future agents 5:16 AI MBA Skool community and written tutorials 5:54 Every's compound engineering plugin automates this 6:29 /workflows:plan command walkthrough 7:22 /workflows:review and /workflows:compound commands 8:22 This loop works across tools and agents 9:33 Live demo: Real client email transformation project 10:16 Client context: German veterinary coach 11:03 Goal: Improve email quality from 6/10 to 8/10 12:19 Pre-work: Analyzing PDF book and Excel spreadsheet 13:40 Sub-agents analyze the German source materials 15:45 Token limits and working with large PDFs 18:22 Analysis complete: Extracting actionable content 21:00 Text extraction with parallel sub-agents 23:08 Starting formal /workflows:plan 24:00 Plan mode spins up research sub-agents 25:00 Token usage: Why Max plan is recommended 26:49 Plan complete: Deepen, Review, or Work options 27:49 Running /workflows:work to execute the plan 39:11 Demo conclusion 39:16 Next step: Learn to create your own skills