What problem does it solve?
Reading a research paper deeply requires tracing every claim back to figures, tables, equations, and experiments, separating author statements from your own analysis, and auditing conclusion boundaries. This Skill automates that structured deep-reading process, producing a fixed Sections 01-16 Paper Card with verifiable source pointers instead of a shallow summary or translated abstract.
Core Features & Use Cases
- Evidence-grounded deep reading: Builds an evidence inventory and claim-evidence matrix linking each central claim to specific figures, tables, equations, and PDF pages before drafting the card.
- Three locator modes: Automatically selects page-grounded, structure-grounded, or source-limited mode based on extraction reliability, and never fabricates page numbers when extraction fails.
- Paper-type lenses: Applies methods, discovery, resource, clinical, materials, or review analytical fragments matched to the paper's argument structure.
- Built-in auditing: Ships with prepare_paper.py and audit_paper_card.py scripts that normalize PDF or source-map input and validate structure, locators, and evidence coverage.
- Use Case: A graduate student receives a new arXiv preprint and asks for a Chinese Paper Card; the Skill extracts the PDF, maps experiments to claims, separates author-stated limitations from critical analysis, and proposes testable follow-up research ideas with explicit hypothesis labels.
Quick Start
Use nature-paper-card to deep-read this paper PDF and generate a complete Sections 01-16 Paper Card with experiment-to-claim evidence tracing and testable research ideas.
