Server has 5 tools with basic schemas and descriptions, but significant gaps prevent higher scores. All tools have descriptions (10-30 chars, below the 50-200 char baseline for LLM optimization), and input schemas are present but minimal. Parameter descriptions are sparse or missing context. No output schemas documented. Error handling is generic. Tool naming follows verb_noun convention (positive), but descriptions lack WHEN/WHY guidance and actionable recovery paths. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite all being READ_ONLY. Composition is reasonable (5 focused tools), but descriptions do not guide LLM selection between similar tools (e.g., get_stock_info vs get_realtime_price distinction unclear).

• Expand tool descriptions to 80-150 chars. Example: 'get_realtime_price: Fetch current A-share price for a given stock symbol (e.g., 000001 for PABANK). Returns price, change%, and timestamp. Use this for real-time quotes; use get_stock_info for fundamentals.'
• Add parameter format constraints in descriptions. E.g., symbol: 'A-share stock code (6 digits, e.g., 000001 for PABANK). Do not include exchange prefix (SH/SZ).'
• Document output schemas for each tool. Example for get_realtime_price: '{symbol: string, price: number, change_percent: number, timestamp: ISO8601, volume: number}'
• Add readOnlyHint annotation to all 5 tools in tool definitions to signal retry safety to LLMs.
• Implement structured error responses with error codes and recovery hints. E.g., 'SYMBOL_NOT_FOUND: Stock code 999999 not found. Valid codes are 6-digit A-share symbols (e.g., 000001). Try search_stocks() to find by name.'

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry