- 好文转载：Fetching Title#brw5
- by meta engineer vasumanmoza
Step 1: architecture.md
Open ChatGPT (4o, not o1/o3/o4) and say:
I’m building a [description of your product - the more detailed the better]. Use Next.js for frontend, Supabase for DB + auth.
Give me the full architecture about this project:
- File + folder structure
- What each part does
- Where state lives, how services connect
Format this entire document in markdown.
Save its output as architecture.md and throw it in an empty folder where your project will live.
Step 2: tasks.md
Now say:
Using that architecture, write a granular step-by-step plan to build the MVP.
Each task should:
- Be incredibly small + testable
- Have a clear start + end
- Focus on one concern
I’ll be passing this off to an engineering LLM that will be told to complete one task at a time, allowing me to test in between.
Save it as tasks.md. Again, throw it in the folder.
Step 3: Cursor
You’re an engineer building this codebase.
You've been given `architecture.md` and `tasks.md`.
- Read both carefully. There should be no ambiguity about what we’re building.
- Follow `tasks.md` and complete one task at a time.
- After each task, stop. I’ll test it. If it works, commit to GitHub and move to the next.
Include this as well - this is crucial:
CODING PROTOCOL
Coding Instructions
- Write the absolute minimum code required
- No sweeping changes
- No unrelated edits - focus on just the task you're on
- Make code precise, modular, testable
- Don’t break existing functionality
- If I need to do anything (e.g. Supabase/AWS config), tell me clearly
This system fixes the biggest problem with vibe coding:
You’re not dumping everything into the IDE and praying. You’re giving it a roadmap. You’re keeping it on rails. You stay in control.
This workflow lets you ship clean, testable AI-assisted code - without the spiral.
Normally I'd ask you to follow me for the playbook but this is literally it. Good luck
← Previous post常用Linux命令