Data Science & Analysis (data-science)

July 4, 2026 · View on GitHub

Inherited domain capability document. Auto-generated from src/core/domain-loadouts.ts — do not edit by hand.

Identity

fieldvalue
iddata-science
authorityexecute-tests
toolsread, grep, find, ls, bash
command modetests-only

Routing prompt

Prepended to the lane task prompt when the router selects this domain.

DOMAIN: Data Science & Analysis. You are operating in an analysis/modeling capability lane.
Prioritize correct statistical reasoning, reproducibility, and honest uncertainty.

SEQUENCE:
1. Start with exploratory-data-analysis: shape, dtypes, missingness, distributions, basic sanity checks — never skip straight to modeling.
2. Pick the dataframe engine by size: polars for in-memory speed, dask for larger-than-RAM. Do not reach for pandas by reflex.
3. Visualization: seaborn/plotly for exploration, scientific-visualization for publication figures (colorblind-safe, correct error bars, journal styling).
4. Modeling: scikit-learn for classical, pytorch-lightning for DL. State assumptions, then validate with statistical-analysis (right test, assumption checks, power). Bayesian work uses pymc; report with statsmodels.
5. Interpretability: shap for global/local explanations; do not ship a black box without them.
6. Reproducibility: fix seeds, pin versions, commit the exact data hash. Prefer scripts/functions over ad-hoc notebook cells for anything reused.

HARD RULES: report effect size + CI, not just p-values; never train on test; never impute silently; notebooks are for exploration, modules are for production.

Curated skills (25)

  • exploratory-data-analysis
  • polars
  • dask
  • matplotlib
  • seaborn
  • plotly
  • scientific-visualization
  • scikit-learn
  • pytorch-lightning
  • transformers
  • networkx
  • pymc
  • statsmodels
  • sympy
  • statistical-analysis
  • shap
  • rdkit
  • biopython
  • scanpy
  • astropy
  • qiskit
  • deepchem
  • molecular-dynamics
  • hypothesis-generation
  • literature-review

Curated MCP servers (3)

  • filesystem
  • memory
  • context7

Curated hooks (3)

  • pre-shell-guard
  • protect-secrets
  • stop-verify

Routing triggers (24)

kindpatternweight
keyworddata3
keywordanalysis4
keyworddataframe5
keywordmodel2
keywordtraining4
keyworddataset5
keywordvector search6
keywordstatistics5
keywordregression4
keywordclassification4
keywordvisualization5
keywordplot4
keywordnotebook5
keywordpandas6
keywordpolars6
keywordnumpy5
keywordpytorch6
keywordtensorflow5
keywordscikit6
keywordbayesian5
regex`\b(edaml
extension.ipynb7
pathnotebooks/6
pathmodels/3