Single tool 'double_number' has a valid schema and description, but the description is in Japanese (non-English), extremely brief (8 chars in English: 'double'), and lacks context for LLM selection. The schema is present but minimal (single numeric parameter). No error handling, no output schema documentation, no composition guidance. Tool naming is acceptable (verb_noun form) but the overall definition quality is well below production baseline. Rubric baseline: A+ tools average 194 chars for descriptions and 100% have param descriptions; this tool has ~8 chars and minimal context.

• Rewrite tool description in English with 50-200 chars explaining WHAT it does, WHEN to use it, and WHAT it returns. Example: 'Multiplies a number by 2. Use this to double numeric values. Returns the doubled result as a number.'
• Add output schema documentation. Declare that the tool returns a structured object with a 'result' field (number) or similar, so LLMs know what to expect.
• Add numeric constraints to the 'num' parameter description. Example: 'A number to double (range: -1e10 to 1e10, no NaN or Infinity).'
• Implement error handling with actionable messages. If input is invalid, return 'Invalid input: num must be a finite number' instead of a generic error.
• Add a tool annotation (readOnlyHint: true) to signal this is a safe, read-only operation that can be retried without side effects.