ds-crew

Use when setting up or repairing the Python env for analysis — detects uv/venv/conda/poetry/pipenv, installs core packages

Use when building or improving a predictive model — AIDE-style solution tree with leakage discipline and empirical leaderboard

Use when a data question is fuzzy or high-stakes — clarifies scope and writes analysis-spec.md before running a solver

Use when starting fresh with data and unsure which skills to use — peeks at data, asks targeted questions, assembles a crew plan

"Use when onboarding a dataset or asking 'what\'s in this data?' — per-column quality report with join-compatibility checks; flags possible PII/sensitive columns"

Use when browsing, pruning, or seeding a run from past analyses stored across sessions

Use when a single hard task keeps failing greedily — tree-searches alternative solution paths via MCTS mode

Use when a number must be right — runs N diverse data scientists in parallel and reconciles into consensus + minority report

Use when answering questions from data files (CSV/JSON/Excel/SQLite) — verified iterative solver with rubric-graded judge and per-role model routing

Use when you already have multiple analysis answers — clusters them into consensus + minority report

Use when analyzing data files with Python — baseline DS-STAR iterative loop with LLM-as-judge; prefer ds-star-plus for production work

Use when auditing any analysis result — grades it against the 6 DS failure modes and returns a score 1–4

Use when checking answer stability — runs the same solver N times and returns the majority answer + agreement rate

Use when exploring data without a fixed question — produces a stakeholder-ready narrative backed by numbers and charts