I just spent a ton of hours developing the best cursor and Google Gemini 2.5 Pro workflow, so you don't have to.

I just spent a ton of hours developing the best cursor and Google Gemini 2.5 Pro workflow, so you don't have to. You can check it out here at this lovely URL. You'll see it's got that cursor rules file up in there and it's also got a whole AI documentation folder with some other interesting bits. And I built this full stack collaborative envelope budgeting app while working on it, so you also get that too in the repo for free. And what I found that really works is starting with a good template. In this case, I started with Wasp, the full stack web app framework for React, Node.js and Prisma, and I used Chad CN admin on top to get all the nice UI components. And you want to use a battle-tested rules for AI file, which I have in the repo. You want to create and constantly revise your plan for implementation. Ask the LLM to look for holes in the logic and your implementation plan and keep improving it before you start. Through doing that, the LLM helped me land on the vertical slice implementation method, which is implementing features that go through the entire stack from UI to database layer, one by one, and then building upon those layers with of complexity as you go along. So yeah, use the vertical slice pattern when building full stack web apps with LLMs. And really key is always ask the LLM to test your assumptions. Also ask it to think of three to four different implementation strategies and then suggest the best one to you before you go ahead and start implementing features. So yeah, really excited to see what other people think about this approach, using this template, using these cursor rules file. Thank you.