hiveguard datasets

Dataset commands require dashboard credentials (HIVEGUARD_DASHBOARD_TOKEN).

datasets list

hiveguard datasets list
hiveguard datasets list --output json

datasets create

hiveguard datasets create NAME [OPTIONS]

Option	Type	Description
--modality	string	image, text, or audio (default: image)
--challenge-type	string	Default challenge type
--prompt	string	Default prompt for items

Example:

hiveguard datasets create "ImageNet Cats" --modality image --prompt "Is there a cat?"

datasets show

hiveguard datasets show DATASET_ID

Shows: name, modality, item count, label count, creation date.

datasets delete

hiveguard datasets delete DATASET_ID

Prompts for confirmation. Deletes the dataset and all its items and labels. Cannot be undone.

To skip confirmation in scripts:

echo "y" | hiveguard datasets delete DATASET_ID

datasets export

Stream a dataset’s labels to a file.

hiveguard datasets export DATASET_ID --fmt csv --output labels.csv
hiveguard datasets export DATASET_ID --fmt jsonl --output labels.jsonl

Option	Type	Description
--fmt	string	csv, json, or jsonl (default: json)
--output	path	Output file (default: stdout)

The export streams directly from the server — safe for large datasets. A .tmp file is written and renamed atomically on completion. If the stream fails, the partial file is cleaned up.