Two tools with basic schemas and minimal descriptions. Both tools have input schemas with proper types (number, string enum), but descriptions are extremely short (10-40 chars), falling well below the 50-200 char baseline for LLM-optimized descriptions. No output schema documentation. Error handling is minimal (divide-by-zero check in 'calculate' returns plain text, not structured guidance). Tool names lack action verbs ('add' and 'calculate' are weak; 'perform_addition' and 'perform_calculation' would be clearer). No parameter descriptions beyond type hints. Schemas are visible and properly typed, preventing a lower score, but the overall quality is below median for production tools.

• Expand tool descriptions to 50-200 chars. Example for 'add': 'Adds two numbers and returns the sum. Use this for simple arithmetic when you need to combine two numeric values.'
• Add parameter descriptions in the Zod schema or via a separate metadata object. Example: 'a: First number to add (any real number)', 'b: Second number to add (any real number)'.
• Document the output schema. Both tools return {content: [{type: 'text', text: string}]}. Explicitly state this so LLMs know what to expect.
• Rename tools to verb_noun format: 'perform_addition' and 'perform_arithmetic' or 'compute_sum' and 'compute_arithmetic'. This makes intent clearer.
• Improve error handling in 'calculate': return structured errors with guidance. Instead of plain text, return {error: 'division_by_zero', message: 'Cannot divide by zero. Provide a non-zero divisor.', recovery: 'Retry with b != 0'}.