Download a workspace folder and everything under it in one session.

Each file lands at / with the workspace's modification time.

A file that already exists locally with the same size and modification time is skipped, so a second run only fetches what changed. Reads run concurrently (-j) and a read the service turns away as busy is retried a few times; one bad file never stops the rest.

Failures are listed at the end and the command exits non-zero if any remain.

With --zip the service packs the folder into one archive and streams it down, which is much faster for large trees.

The archive is extracted into -o with the same layout.

--zip-file keeps the archive itself instead of extracting it, and implies --zip.

The server sends the whole folder every time, so the incremental skip does not apply.

Entries the server left out are listed as SKIP lines; if the archive hit the server's size cap the command says so and exits non-zero.

Examples:

Pull a guide folder

mutiro user workspace pull amelia learning/paxos-guide -o ./paxos-guide

Pull the whole section with fewer parallel reads

mutiro user workspace pull amelia "" -o ./amelia-other -j 2

Pull a large tree as one archive, or keep the archive itself

mutiro user workspace pull amelia registry -o ./amelia --zip mutiro user workspace pull amelia registry --zip-file registry.zip

mutiro user workspace pull <agent> <path> -o <local-dir> [flags]