Planner mode in Cursor is SUPER helpful! With this prompt, agent mode will… ✅ Ask 4-6 clarifying questions ✅ Create a plan with multiple phases ✅ Ask for feedback + approval on the plan ✅ Implement each phase of the plan Steal this prompt down below ⬇️

Add this to your global rules for AI: When asked to enter "Planner Mode" deeply reflect upon the changes being asked and analyze existing code to map the full scope of changes needed. Before proposing a plan, ask 4-6 clarifying questions based on your findings. Once answered, draft a comprehensive plan of action and ask me for approval on that plan. Once approved, implement all steps in that plan. After completing each phase/step, mention what was just completed and what the next steps are + phases remaining after these steps

Just to clarify - this is not a feature within Cursor (yet). This is just a prompt I've come up with called "Planner Mode"

Switch to SkyTab and get $5,000! A modern and sleek POS system with commercial-grade durability. 💪 ✅ $0 upfront costs ✅ Best in-class POS ✅ Local service & 24/7 support ✅ And much more! Make the switch today:

I've been using it for the past day now and is excellent! Was struggling getting the llm to adjust some pieces of code and then entered planner mode and fix it first shot!

Nice!! Love to hear it 💯 You could probably make it even better if you added additional rules to guide planner mode. So in certain folders like my nextjs API folder, I could add a rule there that tells planner mode to structure or name files in a certain way. Lots of ways to play around with it! The example in this video is pretty simple too… will be experimenting with this more for full stack dev (db schema, api dev, frontend dev)

Thanks for sharing your approach. Do you put your frontend and backend code in the same repo? Just curious Also I realized that cursor sometimes doesn't follow rules you give it. Do you have some hacks around that?

1. Best to keep it in the same repo or bring them into the same Cursor workspace - that way the AI can reason about everything across the stack. Adding a rule that tells the AI where your DB schema is and to always read it before making a new API is super helpful. 2. think of some global rules as modes/workflows. my original prompt for cursor to debug in a certain way didn't work. so i said, "When asked to enter Debugger Mode, do xyz..." and that made a huge difference! 2b. be specific with your rules if needed / if they're not a workflow and add them as .mdc rules instead of a global rule

Is this an updated version?

Nope, just a prompt I made which makes Cursor a bit smarter!

Amazing, saved to bookmarks. It is little more advanced, I just started working with prompts and agent mode.

Nice! Agent mode is THE BEST - especially when paired up with a good prompt :)