Three simple, read-only tools with clear naming (pluralize, singularize, is_plural) and complete input schemas. Descriptions are present but minimal (10 - 50 chars), below the 50 - 200 char baseline for LLM-optimized docs. Parameter descriptions are entirely absent, the schema declares 'word', 'count', 'inclusive' but does not explain what each controls or its constraints. No output schema documented. Error handling is basic (generic catch-all). Tools are well-composed (each does one thing) and idempotent, but lack the depth expected of production tools.

• Add parameter descriptions to each tool's inputSchema. For 'pluralize': describe 'word' (the English word to pluralize), 'count' (optional integer; returns singular if 1, plural otherwise), 'inclusive' (optional boolean; if true, prefix result with count). Similar detail for 'singularize' and 'is_plural'.
• Expand tool descriptions to 80 - 150 chars. Example for 'singularize': 'Return the singular form of an English word. Handles irregular plurals (e.g., mice → mouse, children → child). Use this to normalize plural nouns to their base form.'
• Document output schemas in tool descriptions or as a separate schema field. Example: 'pluralize returns {result: string}; singularize returns {result: string}; is_plural returns {is_plural: boolean}'.
• Enhance error messages with recovery guidance. Instead of 'pluralize failed: word must be a string', return 'Invalid word parameter: must be a non-empty string. Example: pluralize(word="cats").'
• Add tool annotations to the TOOLS array: set readOnlyHint=true and idempotentHint=true for all three tools to signal to clients that these are safe to cache and retry.