Three tools with basic schemas and minimal descriptions. All tools have input schemas with proper types (numbers, enums), but descriptions are extremely short (10-40 chars), falling well below the 50-200 char baseline for LLM-optimized descriptions. No output schemas documented. Parameter descriptions are present but terse. No error handling guidance. Tool names lack action verbs (add, mult, calculate are acceptable but generic). No tool annotations (readOnlyHint, destructiveHint). Divide-by-zero error handling exists in calculate() but not surfaced in schema or description.

• Expand tool descriptions to 50-150 chars. Example: 'add: Adds two numbers and returns the sum. Use when you need basic arithmetic addition. Returns a numeric result.' This clarifies intent and distinguishes from calculate().
• Document output schema in tool definitions. Declare that all tools return {content: [{type: 'text', text: string}]} with the numeric result as a string.
• Add readOnlyHint: true to all three tools in their definitions to signal they are safe, idempotent, and have no side effects.
• Replace divide-by-zero text error with a structured error response: {isError: true, content: [{type: 'text', text: 'Division by zero is undefined. Provide a non-zero divisor.'}]} or use MCP error protocol.
• Add numeric constraints to parameter descriptions. Example: 'a: A number (any finite value). b: A number (any finite value except 0 for divide operation).'