Генерация исследовательских гипотез и стратегий

research

research-ideation is a Claude Code skill that generates structured research questions, testable hypotheses, and candidate empirical strategies from a topic, phenomenon, or dataset description in a single pass. Given one input argument, it produces 3–5 questions ordered from descriptive to policy-level, tagging each with a paper type (reduced-form, structural, descriptive, survey-experiment, and others), an identification strategy (DiD, IV, RDD, synthetic control), data requirements, key assumptions, and threats to validity. Output is saved to `quality_reports/research_ideation_*.md`. By default the report goes through a mandatory Post-Flight Verification step: a `claim-verifier` subagent checks negative-literature claims, dataset field assertions, and estimator feasibility against Google Scholar, NBER, and dataset codebooks. The skill is designed for one-shot idea generation; use `/interview-me` to refine a chosen direction further.