How to initialize my AI project

Support Help
1

Hi Team,

I’ve been battling this a couple of days and hope for some advise and guidance.

Goal:
Get LLM/AS/Agents to scaffold the project in multiple platforms (ChatGPT Codex, Claude Code, Google AI Studio and Cursor) simultaneously.

Task:

  • I would like to apply my current engineering rules (~20 files) to an empty project.
  • Then I would like to add the product description and user-stories (~20 files).
  • Then I would like to get the LLM to start building, preferably working through the features/phases them selves.
  • If a correction is needed, I want to add this correction to the “documentation” so I can re-run the project in another instance and also continue with the current project.
  • I what the project to have tracking of all corrections within the repo.

Questions:
Would it make sense to create a development plan with phases?
How do I structure the files/foldes to make the project LLM agnostic?
I’m having a hard time getting the Agents to follow my rules and constantly have to repeat rules.
Is vertical slicing still a thing? :slight_smile:
Am I doing it totally wrong? :slight_smile:

Thanks in advance.

/Simon

2

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.