Use this skill whenever the immediate job is to identify, rank, or normalize Zotero metadata candidates before any create, update, attach, note, or collection write. Trigger even if the user mentions Zotero, as long as they only have a title, DOI, ISBN, or one local PDF/EPUB and need metadata resolved first rather than written now.
This skill resolves metadata candidates before any Zotero write step.
It combines two paths:
It does not create items, attach files, move collections, or write notes.
Use this skill when the immediate task is one of these:
Do not use this skill when:
Prefer one of these entry shapes:
--attachment-path <file>--doi <doi>--isbn <isbn>--title <title> with optional --author <name>--item-type-hint book|journalArticle|documentAlways return structured JSON with:
modequeryattachment_pathitem_type_hintcandidatesEach candidate should contain as many of these fields as available:
item_typetitleauthorsyearpublisherplaceisbndoiurllanguageconfidenceevidencesourcesource_idScript path:
scripts/resolve_metadata.pyExamples:
python skills/zotero-metadata/scripts/resolve_metadata.py `
--doi "10.5555/attention"
python skills/zotero-metadata/scripts/resolve_metadata.py `
--attachment-path "./sample.epub"
After metadata is resolved:
zotero-attachment-resolvezotero-workflow-orchestrator or directly to the local Zotero MCP route