Like all rigging decisions, it depends on what kind of movement and customization you need.
- The easiest way is to make the hand part of the sword image.
- If you want to be able to have multiple hands, then you'd need at least 1) sword, 2) hand behind, and 3) hand in front images.
- If you don't need to move individual fingers, don't waste effort and performance giving each finger bones and images. That could make sense if you actually need it, of course.
Many, many projects have too many details in places they don't need it. Keep in mind the movements you actually need. Don't add rigging for things you don't need.
Similarly, consider how a mesh needs to bend and especially what size your skeleton is viewed at when you are deciding how many vertices a mesh needs. You can make every detail perfect at 10x zoom, but should you? Usually you are doing the work for users who will view it at 1x. Better to spend the time and effort where the user can appreciate it. 🙂 This is not aimed at you specifically, it's just related to making rigging decisions. We see a lot of projects with way more vertices than necessary!
