The concept or "world"., "model" and "local" coordinates are all a bit flexible with scene graphs like the OpenSceneGraph and VulkanSceneGraph, essentially the transform nodes placed in the scene graph hierarchy create a series of local coordinate frame nested within the parents coordinate frame, The concept of "world" and "model" coordinate frames might be treated the same but this is a convention that different development teams may have, it's not something dictated to you by the scene graphs.
If you scene graph has no transform nodes that all the geometry will effectively be in a "world" coordinate frame, if you have just a single transform above your geometry then the local subgraph will be "object" coordinate frame, and the transform above it will place it in the "world" coordinate frame.
I say this all loosely though, some users might be working on visualizing molecules others planetary systems, the scene graphs don't care or dictate what transforms you use or what convention you wish to apply to them. It's the application that decide the meaning of the different coordinate frames.
The .vsgt file you provided looks to be from a .osgb paged database, but isn't a complete conversion as it still refers to .osgb children. vsgconv has support for converting paged databases so you can do the whole lot if you want. Once you've figured everything out it'll be best to convert to .vsgb rather than .vsgt as it'll be smaller and much faster for loading.
The .vsgt has one geometry tile (the VertexIndexDraw node) but it's not decorated by a transform so essentially the data is all in the same coordinate frame. VirtualPlantetBuilder will decorate the tiles with a MatrixTransform to place it from "local" into "world" coordinate frame, but as it's not there I can only presume that the paged database was created by another tool.
So how was this data created?
How familiar are you with the OSG?
Cheers,
Robert.