hey @maxfahl, i continue using a slightly modified version of your prompt.
i made the final addition to the instruction in the beginning section and it saves one precious request for me, maybe you’d like to add too:
You must reply with “I understand the above instructions” when you have read and fully understood the above instructions, reply with:
A summary of what exact steps you will take after reading the above objectives.
Repeat what the execution protocol entails, step-by-step, mark the “Execution Protocol” step as extra important in the summary.
Lastly, if [YOLO MODE] is set to “ask” in “User Input”, ask the user if they want to proceed in “YOLO MODE” or not. Otherwise, respect the “on|off” preference.
“> Perform Execution Protocol Step 1 and Step 2 along with your response.”
I disagree… I have a section in my cursor rules that says, start every response with Hey!.. and as soon as I see it not responding with that, I know context of rules has been lost… And it’s normally very deep in chat session.
@maxfahl This is next level god mode! Thank you! - can i ask, how would you start a new composer session after using this prompt in the previous composer? - I hope that makes sense?
edit - do you use any of the cursor rules or rules for AI with this?
The .cursorrules file is being deprecated in favor of Project Rules (in the .cursor/rules directory). Project Rules give you more control and flexibility - they can match specific file patterns and include semantic descriptions.
For now though, changes to .cursorrules should take effect without needing to restart your composer/chat. The file contents is sent with every query you make to the AI you choose.