The Agent Skills Directory

[SAFE]: The skill's primary entry point, scripts/series_to_volume.py, performs legitimate medical imaging tasks: reading DICOM files, sorting slices, and writing NIfTI volumes. It utilizes well-known, industry-standard libraries such as pydicom and nibabel. All file operations are scoped to caller-provided directories or output paths specified via command-line arguments.
[SAFE]: Data processing is limited to local DICOM headers and pixel data. Metadata extraction (StudyInstanceUID, SeriesInstanceUID, etc.) is performed for technical logging and verification purposes as defined in the validators/output_schema.json. No sensitive patient health information (PHI) is transmitted externally.
[SAFE]: The skill manifest and documentation include clear usage restrictions, explicitly stating that it is not for clinical use or autonomous diagnosis. It relies on standard package managers (pip) for its dependencies (pydicom, nibabel, numpy, typer).
[SAFE]: The fixtures/generate_fixtures.py script creates synthetic data for testing, ensuring no real medical data is required for engineering verification. The security profile is consistent with a developer tool for medical imaging pipelines.

dicom-series-to-volume