Server defines 7 tools with complete input schemas and descriptions in Korean. Naming follows verb_noun pattern (search_memory, save_memory, get_memory, etc.). Descriptions are present but vary in clarity and LLM-optimization. Most parameters have type definitions and descriptions. However, several parameters lack detailed constraints (e.g., limit ranges not enforced in schema, importance scale not documented in parameter description). Output schemas are not explicitly documented, responses are JSON strings without declared field structures. Error handling is minimal; most errors are generic. No tool annotations (readOnlyHint, destructiveHint) despite clear read/write distinctions.

• Add explicit outputSchema declarations to all tools. Document field names, types, and meanings. E.g., search_memory should declare {results: [{id: number, type: string, importance: number, snippet?: string, content?: string}], shown: number, total_matches: number, note?: string}.
• Add minValue, maxValue, and pattern constraints to inputSchema for numeric and enum parameters. E.g., limit: {type: 'number', minimum: 1, maximum: 50}, importance: {type: 'number', minimum: 1, maximum: 5}.
• Add tool annotations: mark save_memory and add_relation with destructiveHint: true; mark all read-only tools with readOnlyHint: true. This enables agents to reason about retry safety and side effects.
• Enhance error messages with recovery guidance. E.g., 'Memory ID 999 not found. Try search_memory(query) to find existing memories, or save_memory(content) to create a new one.'
• Provide English descriptions alongside Korean. LLMs trained primarily on English may misinterpret Korean idioms. Use English as the primary description language for maximum compatibility.