A large project organises its rules into subfolders under `.cursor/rules` by area of the codebase. What happens to a rule file sitting two folders deep?

The principle — Cursor finds project rules by scanning the rules folder, so the scan reaches nested subfolders as well as the top level.

Why the key is correct — A file two folders deep is discovered like any other and applies whenever its frontmatter conditions match. The documentation still recommends a flat layout, but that is a preference about managing the set, rather than a limit on discovery.

Why the others are wrong — Believing that only top-level files are read leads people to duplicate a nested rule at the root, which puts the same text into context twice and pays for it on every request. Depth carries no priority weighting either. And nothing needs to reference a nested file, because discovery is by scan rather than by import.

Remember this — Folders organise rules for humans; the scan reaches the whole tree.

Sources — Cursor's Rules documentation.