The server defines 5 tools with reasonable structure and consistent naming patterns. All tools start with action verbs (get_, analyze_) and have descriptions ranging from 95-158 characters, which falls within the baseline range. However, several definition quality issues reduce the score: parameter descriptions lack constraint details (date format patterns are present in schema but not emphasized in descriptions), no output schemas are documented, and error handling is not visible in the tool definitions. The tools all have proper JSON Schema input definitions with type declarations and descriptions, but lack guidance on error recovery, retryability, and what outputs users should expect.

• Add complete output schema documentation for each tool. Example for get_health_summary: {recovery_score: number 0-100, sleep_quality: number 0-100, stress_level: string 'low'|'medium'|'high', trends: array of {date: string ISO8601, metric: number}, recommendations: array of string}
• For analyze_health_trends, add explicit enum constraint to metric parameter in schema: {enum: ['recovery', 'sleep', 'strain']} and note in description: 'Valid metrics: recovery, sleep, or strain'.
• Expand parameter descriptions with validation rules. Example: 'Start date in YYYY-MM-DD format (required, must be before end_date, max 90 days in the past)' instead of just 'Start date in YYYY-MM-DD format'.
• Add error handling to each tool description. Example: 'If user_id is invalid, returns error 'User not found'. Try omitting user_id to use authenticated user, or verify the ID with your account settings.'
• Document expected data availability. Example for analyze_sleep_patterns: 'Requires at least 3 days of sleep data in the date range. If insufficient data, returns partial results with available dates.'
• Add idempotent behavior note to all analyze_* tools: 'This tool reads data only; it is safe to call multiple times with the same parameters.'