What problem does it solve? Zotero's native "Merge Items" keeps only the master record's field values, silently dropping authors, DOIs, abstracts, and URLs that existed only on duplicate copies, and it only detects duplicates of the same item type — so a preprint and its published journal version are never grouped. This Skill prepares duplicates beforehand so the merge loses nothing. ## Core Features & Use Cases - Duplicate Detection: Find candidate duplicates by title search, shared DOI/arXiv ID, or scan an entire collection with union-find clustering that catches truncated or mangled titles. - Metadata Union: Merge the best value per field across all copies — longest author list, longest abstract, non-arXiv DOI, publisher URL, venue, and date — with optional gap-filling from OpenAlex. - Item Type Normalization: Convert all copies to the type of the most-complete record (journalArticle > conferencePaper > bookSection > preprint), tagging orig-type:/orig-date: for lineage so Zotero can group them. - Use Case: After importing papers via browser extensions, run a collection scan to cluster duplicates, preview consolidation with --dry-run, then commit so Zotero's Duplicate Items merge drops nothing. ## Quick Start Ask Claude to find and consolidate the duplicate records for a paper title in your Zotero library, previewing with a dry run before committing the metadata union and type normalization.
