ListRunFiles

The run ID

Path prefix within the run, e.g. "visualization/" or "Stats/foo". The run is a self-contained namespace: paths do not include any bucket-level or workflow-level layout. Round-trips with File.path returned by this RPC. ".." segments are rejected. Optional: when omitted, lists the whole run (the run root).

If true, list all files recursively under the prefix (default: true) If false, list only files and subdirectories at the current level

Maximum number of files to return per page (default: 1000, max: 1000)

Continuation token from previous response (passthrough from storage)

If true, populate File.download_url (a presigned URL) on each returned file. Defaults to false. Set to true only when the caller intends to fetch file content from the response: generating presigned URLs has a per-object cost and the URLs themselves are credentials with a multi-hour lifetime. Use GetRunDownloadCredentials when long-lived bulk access is required.

Lifetime of the presigned download_url, in seconds. Optional; defaults to 1 hour. Only applies when include_download_url = true. Max 7 days.

If true, presign each download_url with S3 ChecksumMode=ENABLED so the object's stored checksum is returned on the GET (in the x-amz-checksum-<algo> response header) and a client can verify download integrity in a single streaming pass. Defaults false; has no effect unless include_download_url = true. When enabled the presigned URL signs the x-amz-checksum-mode request header, so the fetch MUST send x-amz-checksum-mode:ENABLED — a plain GET that omits it fails with HTTP 403. Best-effort: objects stored without a checksum return none.

Applies only to AWS S3-backed storage connections; other storage backends do not presign S3 GETs and silently ignore this field.