alt Hacker NewsI'd been using this workflow for a while, but this post I found on HN a couple of weeks ago really solidified it: https://boristane.com/blog/how-i-use-claude-code/
> The workflow I’m going to describe has one core principle: never let Claude write code until you’ve reviewed and approved a written plan. This separation of planning and execution is the single most important thing I do. It prevents wasted effort, keeps me in control of architecture decisions, and produces significantly better results with minimal token usage than jumping straight to code.
Replies
arikrahman • today at 5:08 PM
Manus workflow uses three file approach via deliverables.md, taskplan.md, and notes.md. I use this combined with VSDD with my agents.
swaminarayan • today at 5:03 PM
sure will take a look
I’ve been following that post too since I started using Claude (about two weeks now) and it’s great. Sometimes for small changes I shorten research+plan into one step but I always tell it to stop and wait before it writes the code.
One thing I’ve learned: if you notice the agent spinning its wheels, then throw the work away, identify a fix (usually to Claude.md) and start over. Context rot is real.