Scale your data workflows beyond available memory.
Guide parallel and distributed Python computations so users can process data larger than available memory.
When to use it
Use when datasets exceed RAM, pandas or NumPy operations need scaling, or custom computations need parallel execution with dependencies.
It changes how the agent implements data computations by selecting Dask collections and schedulers appropriate to the project's data and operations.
What you provide
Requires Python 3.10+; the documented Zarr 3.4 integration needs Python 3.12+.
Requires dask 2026.8.0.
This skill
https://arxiv.org/abs/2609.00065
Fetches the author's paper record
dask
Installs Dask array and dataframe support