[list] Copying large folders is unreliable

We have several issues that creates instability with large (> 500 items) folder copies.

1. Fundamental issues with content integrity we (or S3 SDK) dies on - "empty" or `//` prefixes
2. The relationship with `DA_JOBS` seems brittle. I cannot reliably copy anything of meaningful size.
3. We are not resilient to "Network changed" errors that may happen on the client. Ideally the client has a way to pick up where they left off. This could be as simple as getting the continuation token and when supplying back da-admin looks at DA_JOBS to see if there's an existing job.

No files, tests, or entry points are named. Start by tracing large-folder copy handling in da-admin, including the S3 SDK interaction and DA_JOBS relationship; reproduce a copy over 500 items and document what completion and resume behavior should look like.

Written by the indexing model from the issue text.