Two tools with minimal descriptions (both under 20 chars in English context, Japanese descriptions present but not evaluable for LLM optimization). Input schemas present but sparse, only 'url' and 'sheetName' parameters with basic string types, no enums, ranges, or validation hints. No output schemas documented. No error handling guidance. Tool names are verb-based (getSpreadsheet, getSheetData) but descriptions lack WHAT/WHEN/WHY context needed for LLM selection. No parameter descriptions visible in schema definitions. Missing idempotency hints, pagination support, and field-level documentation.

• Rewrite tool descriptions in English, 50-200 characters, answering: What does it do? When use it instead of alternatives? Example: 'getSpreadsheet: Fetch metadata (name, sheet list, last modified) for a Google Sheet. Call this first to discover available sheets before reading data.'
• Add parameter descriptions to schemas: 'url' → 'Full Google Sheets URL (https://docs.google.com/spreadsheets/d/...) or spreadsheet ID'; 'sheetName' → 'Exact sheet tab name (case-sensitive). Call getSpreadsheet first to list available sheets.'
• Document output schemas inline or in README. Example: getSpreadsheet returns {id, name, url, lastModified, sheets: [{id, name, index, rowCount, columnCount}]}. getSheetData returns 2D array of mixed types.
• Add error recovery hints: 'If sheet not found, call getSpreadsheet to verify sheet name. If API fails, check GAS_WEB_APP_URL and GAS_API_KEY environment variables are set.'
• Add pagination to getSheetData: accept optional 'maxRows' (1-10000, default 100) and 'startRow' (0-indexed) parameters. Return {data: [[...]], totalRows, hasMore} to enable chunked reads.