Server defines 7 tools with complete input schemas and descriptions. All tools follow verb_noun naming (search_*). Descriptions are present and moderately detailed (100-250 chars), explaining purpose and parameters. However, descriptions lack actionable guidance on when to use each tool vs. alternatives, and output schemas are documented only in docstrings, not in structured form. Parameter descriptions are adequate but generic. No error handling guidance, no examples of recovery paths. Tools are well-composed (each does one thing), but descriptions could better explain the financial product taxonomy and when to call search_financial_company first.

• Add explicit JSON Schema definitions for tool outputs. Document the structure of 'products' array, 'total' count, 'page', 'size', 'kind', 'raw_endpoint', and optional 'groups' field. This enables LLMs to extract and chain results.
• Enhance descriptions with dependency hints: 'Call search_financial_company first to find company_code if you only have a company name. Then pass company_code to this tool for precise results.'
• Define top_group as an enum with all valid 권역코드 values (020000, 030300, 050000, 060000, etc.) and their meanings (은행, 저축은행, 보험, 금융투자). Replace free-form string with constrained input.
• Add error handling guidance to all tool descriptions: 'If no results found, try broadening the query or omitting company_code. If API returns 429, retry after 60 seconds.' This guides LLM recovery.
• Provide English translations or explanations for Korean financial terms in descriptions (e.g., '적금 = savings account', '정기예금 = fixed deposit'). This helps non-Korean-speaking LLMs understand the domain.