Two tools with minimal descriptions and no visible input schemas. studentDetails has a 127-char description (acceptable range) but lacks parameter type definitions in the visible schema. listStudentIds has a 56-char description (below 20-char floor for scoring). Neither tool has documented output schemas. No error handling guidance, no parameter constraints, no composition hints. The Tools class uses Spring AI annotations (@Tool, @ToolParam) but the actual JSON Schema registration is not visible in the provided source, tool definitions appear inferred from annotations rather than explicitly registered with full schemas.

• Add explicit JSON Schema definitions for both tools. For studentDetails, declare studentId as type 'string' with minLength, pattern, or enum constraints. For listStudentIds, declare empty input schema with type 'object' and no required properties.
• Expand listStudentIds description to 100-150 characters explaining what the comma-separated format contains (e.g., 'Returns a comma-separated list of all valid student IDs. Call this first to discover available students before fetching details.').
• Document output schemas. Specify that studentDetails returns a JSON object with fields (name, age, GPA, etc.) or a plain string; specify that listStudentIds returns a single comma-separated string of IDs.
• Add error recovery hints to studentDetails description: 'If student ID is not found, call listStudentIds() to see valid IDs.' This guides the LLM to self-correct without trial-and-error.
• Add parameter constraints to studentDetails: 'studentId must be a non-empty string matching the format returned by listStudentIds().' This prevents invalid calls and reduces error handling burden.