If no one else is doing this, I'll work on a PR for this.

I was under the impression that in order to get a full snapshot, you had to use ListChangedBlocks and iterate each snapshot to account for all the blocks. However, after playing with the API, I realized that ListSnapshotBlocks truly returns all the blocks for a snapshot and missing blocks appear to be just unused blocks and thus can just be empty. When I talked to a couple colleagues both at AWS and outside that are much more well versed than I at the API, they were also under the impression that this was the case. When I saw "Incremental snapshot support" I also assumed that this was the need. Now knowing that ListSnapshotBlocks does indeed return the complete set of snapshots blocks, I no longer need to implement the diffing which I assumed to be "incremental support". Not sure if that was what you mean by "Incremental snapshot support".

That being said, I created https://github.com/forensicmatt/aws-snap-io to implement Read/Seek around Snapshots in a rudimentary way. Planning to use this for some projects I will soon publish so that volume/file system access based off of libraries that can handle Read/Seek traits can be used.