Three tools with basic schemas and minimal descriptions. All tools have input schemas with proper types (Zod validation), but descriptions are extremely short (10-34 chars, well below the 50-200 char LLM-optimized baseline). Parameter descriptions are present but terse. No output schema documentation. Error handling is minimal (only divide-by-zero check in 'calculate'). Tool names lack action verbs and are generic. No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Composition is reasonable (single responsibility), but descriptions do not guide LLM selection or explain when to use each tool vs. alternatives.

• Expand tool descriptions to 50-200 chars. Example for 'add': 'Performs simple addition of two numbers. Use this for basic arithmetic when you need to add two values together. Returns the sum as a number.'
• Rename tools with action verbs: 'add' → 'perform_addition', 'calculate' → 'perform_calculation', 'calculate-bmi' → 'compute_bmi'. This makes intent clear from the name alone.
• Document output schemas explicitly. Add to each tool description: 'Returns: {content: [{type: "text", text: string}]}' or similar structured format.
• Add tool annotations to the server.tool() calls. Use readOnlyHint: true and idempotentHint: true for all three tools to signal they are safe to retry.
• Add input validation and error messages. For 'calculate-bmi', check heightM > 0 and weightKg > 0; return 'Error: height must be > 0 meters' if invalid. For 'calculate', validate a and b are numbers.