Server has 10 tools with complete input schemas and descriptions. Naming follows verb_noun convention (search_pages, get_page, create_page, etc.). Descriptions are present but brief (avg ~50 chars, below the 194-char baseline). Schemas are well-structured with proper types and enums for block types. However, output schemas are not documented, LLMs cannot predict response structure. Error handling exists but lacks recovery guidance. Security validation is implemented (notionIdSchema, safeTitleSchema) but not exposed in descriptions. No tool annotations (readOnlyHint/destructiveHint) despite clear read/write distinction.

• Document output schemas for all tools. Example: search_pages returns {pages: [{id, title, url, created_time}], total_count}. This enables LLMs to chain tools and extract the right fields.
• Expand tool descriptions to 100-200 chars. Example: 'Get basic information about a Notion page' → 'Retrieve page metadata (title, created_time, last_edited_time, icon, cover). Use this to check page properties before editing. For page content/blocks, use get_page_content instead.'
• Add tool annotations to schema. Mark read-only tools with readOnlyHint: true and write tools with destructiveHint: true (for create/update/delete). This tells LLMs which tools are safe to retry.
• Enhance error messages with recovery guidance. Example: 'Page not found. Try search_pages(query=<partial_title>) to find the correct page_id, or verify the ID format is a valid Notion UUID.'
• Add format constraints to parameter descriptions. Example: 'page_id: Notion page ID (32-char hex string, e.g. 550e8400e29b41d4a716446655440000)' or use regex pattern in schema.