The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The documentation provides instructions to install several standard Python libraries from official registries, including zarr, s3fs, gcsfs, numpy, dask, xarray, h5py, and pandas. These are well-known and reputable packages used for scientific computing.
[COMMAND_EXECUTION]: The skill includes shell commands for package management via uv pip and Python code snippets that interact with the local filesystem and cloud storage backends (S3/GCS).
[PROMPT_INJECTION]: The skill documentation identifies surfaces for indirect prompt injection by processing external data stores. Ingestion points: Data enters the context through zarr.open(), zarr.open_array(), and xr.open_zarr() in SKILL.md. Boundary markers: There are no instructions defining delimiters or warnings to ignore embedded instructions within the processed data. Capability inventory: The skill describes capabilities for file writes (zarr.create_array, z[:] = ...) and network access through integrated cloud storage libraries. Sanitization: The documented library usage patterns do not include explicit content validation or sanitization.
[SAFE]: The documentation does not contain any malicious patterns such as direct prompt injection, obfuscated code, hardcoded credentials, or persistence mechanisms. Its primary purpose is to educate on the use of the Zarr library.

zarr-python