This server implements 2 tools with basic HTTP transport via Spring Boot. Both tools have input schemas and descriptions present, but quality is below production baseline. Tool names lack clear action verbs (queryListByName should be get_students_by_name or search_students), descriptions are minimal (under 100 chars), and parameter descriptions are generic. No output schema documentation visible. No error handling guidance. The implementation relies on Spring AI's MethodToolCallbackProvider for tool registration, which is reasonable, but the actual tool definitions lack the LLM-optimization depth needed for reliable agent selection and parameter binding. Compared to the 194-char average description baseline, these tools fall short. No evidence of pagination limits being enforced, permission gates, or structured error responses.

• Rename 'queryListByName' to 'search_students_by_name' (verb_noun format, clear action).
• Rename 'pageInfo' to 'list_students' (shorter, clearer verb_noun) or 'search_students' (if filtering is primary).
• Expand tool descriptions to 50 - 200 chars. For search_students_by_name: 'Search for students by full or partial name. Returns up to 20 matches. Call this first if you only have a student name; use list_students for pagination and filtering.'
• Expand tool descriptions for list_students: 'Retrieve paginated student records with optional filters (name, sex, classroom, address). Returns page count, total count, and chaining IDs (student_id). Useful for discovering students or applying multiple filters.'
• Document parameter constraints in descriptions. For 'sex' param: '(Filter by gender: M for male, F for female. Optional; omit for all.)' or declare as enum: ["M", "F"].
• Document expected pagination defaults in list_students description: 'Defaults: current=1, size=10. Max size=100. Returns: items[], total_count, current_page, has_more.'
• Add output schema documentation in tool config or docstring. E.g., 'Returns: {students: [{id, name, sex, classroom, address, created_at}], total, current, pageSize}'.
• For pageInfo, add per-parameter validation rules: 'current: integer, min=1 (page number), size: integer, 1 - 100 (items per page), name/sex/classRoom/address: optional string filters (exact match or wildcard).'
• Implement error handling with actionable guidance. E.g., if pageInfo receives size=1000, return: 'Page size too large: requested 1000, max is 100. Retry with size=100.'