The Agent Skills Directory

[SAFE]: The skill facilitates the use of the zarr Python library, a standard tool for scientific data management. The instructions involve typical data science workflows such as array creation, indexing, and storage backend configuration.
[EXTERNAL_DOWNLOADS]: The skill recommends installing standard, well-known libraries from official package registries (e.g., pip install zarr, s3fs, gcsfs). These are verified dependencies for the library's functionality.
[INDIRECT_PROMPT_INJECTION]: As a data-processing utility, the skill creates an attack surface where the agent might ingest untrusted data or metadata from external Zarr stores (local or cloud). This is a standard risk for any tool reading external files.
Ingestion points: Loading arrays and metadata via zarr.open_array, zarr.open_consolidated, or xr.open_zarr from potentially untrusted S3/GCS buckets or local directories.
Boundary markers: None present in code snippets.
Capability inventory: File system read/write via LocalStore, network operations via s3fs and gcsfs plugins.
Sanitization: Standard library use; validation of data content is left to the user/application layer.

zarr-python