Server has 4 tools with complete input schemas and descriptions. Naming follows verb_noun convention (analyze_text, generate_outline, format_gbt7714, count_frequency). Descriptions are present and contextual (avg ~150 chars), exceeding the 10-char minimum. All parameters have type definitions and descriptions. However, output schemas are not documented, LLMs cannot predict return structure. Error handling is minimal (no recovery guidance, no categorization). No tool annotations (readOnlyHint, destructiveHint). Parameter descriptions lack format constraints (e.g., 'accessDate' should specify ISO 8601). Tools are read-only but this is not declared via annotations.

• Add output schema documentation to each tool. For analyze_text, document: {editorWordCount: number, chineseChars: number, paragraphCount: number, estimatedReadingTimeMin: number, ...}. Use JSON Schema format in tool definition.
• Add readOnlyHint: true annotation to all 4 tools. Update tool registration to include annotations: {readOnlyHint: true} for each tool definition.
• Enhance parameter schemas with JSON Schema constraints. For format_gbt7714.accessDate, add format: 'date' or pattern: '^\d{4}-\d{2}-\d{2}$'. For count_frequency.topN, add minimum: 1, maximum: 500 in schema.
• Implement error recovery guidance. When text is empty, return: {error: 'Text cannot be empty. Provide at least 1 character.', retryable: true}. Categorize all errors as retryable, user_fixable, or fatal.
• Document parameter relationships. In format_gbt7714, note that 'volume' and 'issue' are only relevant for type='journal'. Use conditional schema or explicit description: 'volume and issue apply only to journal articles.'