What is the difference between adding a new section and using existing code caves when backdooring an EXE?

Adding a new section with RWE permissions (e.g., `.sdata`) is straightforward and doesn't rely on finding large enough code caves, but it increases file size. Using existing code caves keeps the file size unchanged, but the payload must fit within the cave's size limit, and non-executable sections require permission modifications. BDF supports both methods, as shown in the practical testing section.