Skip to content

iowarp/science-data

v0.1.1Apache-2.0

Science data pack: a skill that reads HDF5, NetCDF, Parquet and CSV with Clio's data tool first, and declares the clio-kit parquet and pandas MCP servers with per-tool action classes for analysis beyond it. Read tools run without asking; file-writing tools ask first; each server needs one online launch to build its environment before it runs offline.

GitHub
science-data

Explores HDF5, NetCDF, Parquet and CSV data read-only: Clio's data tool for structure, slices and summary statistics, then the clio-kit parquet and pandas servers for aggregation, correlations and statistical tests, and the hdf5 server or h5py only for what those cannot do. Not for writing or converting data files.