What problem does it solve? Papers often mislabel their own study design, calling retrospective database analyses "real-world evidence" or omics screens "mechanism studies," which distorts literature appraisal and evidence grading. This Skill classifies what a paper actually did based on its structural features rather than author self-description. ## Core Features & Use Cases - Structure-Based Classification: Maps papers to design families (RCT, cohort, case-control, cross-sectional, diagnostic, prognostic, omics, mechanistic) using population, allocation, timing, and comparison logic. - Hybrid Design Separation: Distinguishes primary and secondary design layers in multi-component papers such as TCGA screening plus wet-lab validation. - Mislabel Correction and Confidence Rating: Flags inflated author terminology and assigns High/Medium/Low classification confidence with an evidence-family position for downstream appraisal. - Use Case: A journal club reviewer asks whether a biomarker paper is a true validated prediction study or exploratory omics screening, and receives a design-identification memo with rejected alternatives and confidence level. ## Quick Start Ask the assistant to identify the real study design of a paper by pasting its abstract, methods section, or a structured study summary.