data-science-agent-skills

Convert analysis into stakeholder-ready insight narratives that state decisions, evidence, uncertainty, and recommended action without overclaiming. Use when an agent needs a judgment-heavy data science workflow for convert analysis into decision-grade narrative, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Help teams decide what can be learned responsibly from small, imbalanced, expensive, or scarce datasets. Use when an agent needs a judgment-heavy data science workflow for avoid overfitting scarce data, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Reconstruct practical lineage across SQL, notebooks, scripts, dashboards, model artifacts, and pipeline configs. Use when an agent needs a judgment-heavy data science workflow for rebuild lineage from sql, notebooks, and configs, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Convert data science prototypes into engineering-ready ML handoffs with contracts, artifacts, SLOs, monitoring, and ownership. Use when an agent needs a judgment-heavy data science workflow for turn prototypes into engineering handoffs, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Guide data science work on sensitive data using minimization, local-first analysis, de-identification, synthetic data, or privacy-preserving methods. Use when an agent needs a judgment-heavy data science workflow for choose safe analysis patterns for sensitive data, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Audit open-source packages, notebooks, model weights, datasets, licenses, and provenance used in data science work. Use when an agent needs a judgment-heavy data science workflow for audit ds dependencies, models, data, and licenses, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Convert exploratory notebooks into modular, parameterized scripts or orchestrated pipeline tasks. Use when an agent needs a judgment-heavy data science workflow for refactor notebooks into maintainable pipelines, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Package notebooks so another machine or teammate can reproduce their results with known data, environment, seeds, and order. Use when an agent needs a judgment-heavy data science workflow for make notebooks runnable elsewhere, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Create concise model risk and compliance memos for review, audit, launch, or executive approval. Use when an agent needs a judgment-heavy data science workflow for create governance-ready model risk memos, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Diagnose missing-data mechanisms and design imputation, exclusion, or sensitivity plans. Use when an agent needs a judgment-heavy data science workflow for diagnose missing data before imputation, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Review whether a chosen data science metric actually supports the decision, user outcome, risk, and cost tradeoff. Use when an agent needs a judgment-heavy data science workflow for align metrics with real decisions, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Design transparent, reversible cleaning plans for messy datasets without silently mutating evidence. Use when an agent needs a judgment-heavy data science workflow for design reversible data cleaning plans, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Actively search ML datasets and pipelines for target, temporal, group, split, join, and preprocessing leakage. Use when an agent needs a judgment-heavy data science workflow for attack ml pipelines for leakage, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Decide whether a feature store is needed and design features to avoid training-serving skew, staleness, and reuse confusion. Use when an agent needs a judgment-heavy data science workflow for decide when a feature store is worth it, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Preserve experiment provenance across code, data, environment, parameters, metrics, artifacts, and human decisions. Use when an agent needs a judgment-heavy data science workflow for capture code, data, env, and decision history, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Design evaluation strategies that match task type, deployment setting, time, groups, and decision risk. Use when an agent needs a judgment-heavy data science workflow for design trustworthy model evaluation, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Design drift monitoring for production models and data products across data, concept, performance, and operational signals. Use when an agent needs a judgment-heavy data science workflow for design data, concept, and performance drift monitors, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Structure collaboration between data scientists, AI agents, and domain experts when domain judgment is required. Use when an agent needs a judgment-heavy data science workflow for structure sme review for agent-assisted ds, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Infer practical data contracts, schema expectations, and validation tests from messy real datasets. Use when an agent needs a judgment-heavy data science workflow for infer contracts and tests from messy data, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Run root-cause analysis for broken models, metrics, dashboards, pipelines, and data science decisions. Use when an agent needs a judgment-heavy data science workflow for diagnose broken models and dashboards, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Turn ambiguous business, product, policy, or research asks into decision-ready data science problem statements. Use when an agent needs a judgment-heavy data science workflow for turn vague asks into measurable ds problems, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Plan compute use so data science experiments, model training, and inference stay within budget and latency constraints. Use when an agent needs a judgment-heavy data science workflow for plan gpu, cloud, and sampling costs, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Clarify when a data science request needs causal reasoning, what assumptions are required, and what evidence is missing. Use when an agent needs a judgment-heavy data science workflow for stop predictive work from pretending causality, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Design honest baselines and ablations before complex models are accepted. Use when an agent needs a judgment-heavy data science workflow for force honest baselines and ablations, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.

Assess whether a dataset or data domain is ready for AI, analytics, or modeling before a project commits time and budget. Use when an agent needs a judgment-heavy data science workflow for score whether data is usable for ai work, including evidence review, local artifact inspection, risk classification, stakeholder-ready decisions, reproducibility, governance, or agent-to-agent handoff. Trigger for Codex, Claude, Gemini, Copilot, Cursor, Windsurf, Gravity, LangGraph, CrewAI, AutoGen, or local agents when this exact workflow is needed.