Four tools with clear verb-noun naming (list_, search_, get_, search_in_). All have descriptions (avg 150 chars, within 10-1024 range). Input schemas present with type definitions and parameter descriptions. However, output schemas are not documented, responses are model_dump() calls with no schema visibility. Error handling exists (LibraryError wrapper) but lacks recovery guidance. No tool annotations (readOnlyHint, etc.). Composition is sound: tools chain well (search_papers → get_paper → search_in_paper). Missing: output schema documentation, per-parameter constraints (enums), and actionable error messages.

• Document output schemas for all tools. E.g., list_collections returns {collections: [{name: string, paper_count: integer}]}. Include this in tool descriptions or a separate schema reference.
• Add readOnlyHint: true to all tool schemas to signal to LLMs that these tools are safe to call without side effects.
• Enhance error messages with recovery guidance. E.g., 'Paper not found. Try search_papers() to find the correct paper_id.' instead of bare error code.
• Add enum constraint to 'collection' parameter in search_papers. Populate from list_collections() result or document valid values.
• Document pagination behavior in search_papers and search_in_paper descriptions. State that limit caps results at 10 and explain how to retrieve more (e.g., 'Use limit parameter to fetch up to 100 results per call').

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

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