What problem does it solve? Data analysis failures are silent: code runs clean and returns a confident, wrong answer. This Skill establishes a discipline layer for empirical work in R, Julia, and Python, forcing validation, planning, and human checkpoints before any number reaches a stakeholder. ## Core Features & Use Cases - Skill routing: Maps any analysis request (estimate an effect, clean data, build a figure, verify a number) to the right discipline skill such as question-framing, data-contracts, causal-identification, structural-estimation, or predictive-modeling. - Workflow fork: Distinguishes descriptive, reduced-form causal, structural, and predictive goals so each question is answered with the appropriate method and assumptions. - Durable state and checkpoints: Maintains a docs/analysis/ state folder that survives session resets, and enforces human-in-the-loop gates on design, sample, spec, or estimand changes. - Use Case: A researcher asked to "just run the regression" is routed through executing-analysis-plans with data-contract validation on each step, rather than shipping an unverified estimate. ## Quick Start Ask the assistant to analyze a dataset or estimate an effect, and it will invoke this skill first to frame the question and route to the right discipline before touching the data.
