What problem does it solve? Finding defensible Nature/CNS-family references for manuscript claims is slow and error-prone: you must split text into citable statements, search scholarly databases, verify journal scope, and format exports for reference managers. This Skill automates that pipeline while keeping evidence grading conservative. ## Core Features & Use Cases - Segment-level citation mapping: Splits manuscript text into stable segments (S001, S002, ...), generates English search queries per segment, and maps each to candidate papers from Nature Portfolio, the AAAS Science family, or Cell Press. - Reference-manager exports: Writes a single ENW, RIS, or Zotero RDF file without inventing missing bibliographic fields, plus optional HTML/TSV/JSON review artifacts with year filtering and selective download. - Long-text batch processing: Handles 10+ segment inputs with batch mode, partial checkpoint exports, and automatic retries on transient Crossref failures. - Use Case: Paste a Chinese or English Introduction paragraph, ask for "CNS及子刊引用，导出 Zotero RDF", and receive a segment-to-reference correspondence table plus an importable .rdf file and an HTML browser for manual screening. ## Quick Start Add Nature/CNS citations to this manuscript paragraph, segment by segment, and export the results as a RIS file for EndNote.
