Question about expected behavior of indexing with regards to the context

I’ve just started using Cursor more seriously and I am trying to figure out what is the limits/capabilities of tool. One thing that bothers me is that I am trying to find a quick way to share large areas of the code with the chatbot without having to individually add each and every file.

I’m slowly wrapping my head around the codebase indexing system. Seems that:

  • unless I use @Codebase the chatbot does not see any of my files or their content even though the project is indexed.
  • If I use @Codebase the chatbot gets the files and their contents added to the context

Is that the expected behavior? Or is the bot supposed to see the files indexed in the context even without using @Codebase.

Second question:

I am really struggling to get Cursor to acknowledge anything about a second folder added to the same workspace. If I add a second folder and re-index, the chatbot does not see the contents of that folder regardless of whether I use @Codebase or even @Folder and point to that exact folder.

Can someone please bring some clarification as to what is the expected behavior of the indexing system, and shed some light as to why Cursor struggles so much to handle a second folder?

To reproduce:

Add a folder to the workspace. Chat with the bot using @Codebase to get the files from the first folder added to the context. Then add another folder to the workspace. Re-index. Notice that Cursor does not see or know anything about that new folder.