I am building a prototype to have a single Application providing models for forge viewer. Each user should have its model and not be able to see models of others.
For this I would consider "separation in application architecture” like described here: https://forge.autodesk.com/blog/accounts-apps-keys-and-ids (manual app creation per client is not an option). If bucket are generated in a way that cannot be guessed, we could consider separation per bucket enough (even if not fully secure on “lucky” hit)
I tried to use scope viewables:read, which from documentation, is "only be able to read the end user’s viewable data”. I cannot list buckets (GET buckets), as expected, but I can access bucket by name (GET buckets/:bucketKey/details).
Is viewables:read safe enough for what I described, or can I expect some other points to be readable, but not expected like this one?
Is viewables:read the option discussed in this article https://forge.autodesk.com/blog/securing-your-forge-viewer-token-behind-proxy ? Or is forge roadmap expecting some finer grain access control?
Thank you,