Korean Stock MCP has 5 tools with complete input schemas and descriptions, but quality is uneven. Tool names lack action verbs (stock_price vs get_stock_price), descriptions are adequate but generic (avg ~80 chars, below 194-char baseline), and parameter descriptions are minimal. No output schemas documented. Error handling returns plain strings without recovery guidance. All tools are READ_ONLY, reducing risk, but composition could be tighter (stock_search + stock_price could chain better). Schemas are present and typed, preventing a lower score, but lack depth for production use.

• Rename tools to action-verb form: get_stock_price, get_stock_detail, search_stocks, get_market_index, get_stock_news. This signals intent to LLMs and follows 90% baseline of A+ tools.
• Document output schemas as JSON objects. E.g., stock_price returns {name: string, code: string, price: number, change: number, ratio: number, status: string}. This enables LLMs to chain tools and extract fields reliably.
• Enhance error responses with recovery guidance. Instead of 'Error: 종목코드 {code} 조회 실패', return 'Stock code {code} not found. Try search_stocks(query="company name") to find the correct code.' This guides LLM recovery.
• Add parameter validation rules to descriptions. E.g., 'code: 6-digit Korean stock code (e.g., 005930 for Samsung). Invalid codes return 404.' This prevents LLM hallucination of invalid inputs.
• Add tool dependency hints. E.g., stock_price description: 'Get current price for a stock. If you only have a company name, call search_stocks() first to find the code.' This optimizes agent planning.