Related Servers

Purpose boundaries are unusually well documented: the five search tools (semantic, substring, tag, citation-key, advanced) are explicitly differentiated and cross-referenced, and the content-reading tools (fulltext, read_pdf_pages, outline, page_layout, attachment_path) each target a distinct need. A few pairs (search_items vs advanced_search, get_annotations vs synthesize_annotations, get_notes vs get_annotations) still overlap enough that misselection is possible, so it is not a perfect 5.

All tools use a uniform zotero_ prefix with snake_case verb_noun phrasing (get_, search_, create_, update_, delete_, add_, attach_, set_, list_, switch_, batch_, export_). The only deviation is zotero_write_capabilities, a noun phrase rather than verb_noun, making the set 'mostly consistent with minor deviations' rather than flawless.

41 tools is well beyond the typical 3-15 well-scoped range and exceeds the 25+ 'too many' threshold; several search and content-read tools could reasonably be consolidated. The domain (local+web access, PDFs, semantic index, bibliography) is genuinely broad, which mitigates but does not eliminate the heaviness of the surface.

Coverage is extensive: full CRUD for items, collections, notes, and annotations, plus attachments, multi-mode search, fulltext/PDF reading, semantic indexing, bibliography export, and library switching. Minor gaps remain (standalone notes without a parent, empty-trash/permanent deletion, explicit attachment deletion), but core lifecycles are covered.