OpenAlex Works Explorer provides 14 tools covering scholarly data exploration across works, authors, sources, institutions, topics, publishers, and funders. Tool naming follows the verb_noun pattern consistently (search_*, get_*) which is strong. However, definition quality is hampered by: (1) incomplete parameter descriptions, many parameters like 'filters' and 'sort' lack actionable format guidance; (2) missing output schema documentation, no explicit description of what fields are returned for each tool; (3) generic descriptions for similar tools that blur distinction; (4) no error handling guidance visible in tool definitions; (5) no per-tool result limits documented (critical for search tools). The 'summarize_results' feature in search_works is well-thought-out, but inconsistently applied across other search tools. Overall, the tools are functional and well-named, but lack the depth of parameter constraint documentation and output structure clarity needed for production-grade definition quality.

• Add explicit output schema documentation to every tool. For search_works, document: 'Returns {results: [{id, doi, title, publication_year, authors: [string], cited_by_count, venue, oa_url, abstract}], meta: {count, cursor, per_page}}' (or full object schema if summarize_results=false).
• Specify valid filter keys in search_* tool descriptions. Example: 'filters: Object with keys from {publication_year (int), is_oa (bool), has_abstract (bool), language (string: en|fr|...), type (string: journal|conference|...). Use pipe (|) for OR, exclamation (!) for NOT. See https://docs.openalex.org/api-entities/works/filter-works for full list.'
• Document maximum result limits and pagination behavior: 'Supports up to 200 results per page (per_page parameter). For results beyond the first 200, use cursor pagination, never use offset. Cursor '*' starts from the beginning; the response includes the next_cursor for the next batch.'
• Enhance get_* tool descriptions with entity-specific details. Example for get_work: 'Retrieves complete metadata for a work, including authorships, publication venue, open access status, citation metrics, and abstract. Useful for detailed analysis after search_works identifies relevant papers.'
• Add error recovery guidance to all tool descriptions. Template: 'If [specific error condition], [recommended recovery]. Examples: "If work_id not found, use search_works to find the correct ID." "If cursor invalid, restart with cursor=* to re-fetch from the beginning."'
• Standardize the summarize_results feature across all search_* tools. Either apply to all (recommended for consistency), or document why works differ from authors/sources/etc.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry