This MCP server exposes 6 tools with significant quality gaps across naming, description, and schema documentation. The tools are a heterogeneous collection (skill routing, MCP installation, document conversion, and e-commerce scrapers) with minimal consistency in definition quality. Most critically: (1) Tool descriptions range from moderately detailed to entirely missing context about prerequisites, return types, and error conditions. (2) Input schemas are present for most tools but lack detail, parameter descriptions are minimal or absent in several cases. (3) No output schemas are documented, forcing LLMs to guess at response structure. (4) The 'install_mcp_tools' tool accepts an empty input object, providing no guidance on configuration or secrets. (5) The three scraper tools (shopee_extract, tiktok_shop_extract, wechat_extract) have identical single-parameter schemas with no error guidance, rate-limiting warnings, or response structure documentation. (6) The 'book_to_skill' tool references a Python module that may not have proper error handling or output schema definition visible in the MCP registration. (7) Tool names lack consistency: 'skill_router' uses snake_case, 'install_mcp_tools' is descriptive, but scraper tools lack context about what data structure they return or how to handle failures.

• Add output schemas for all 6 tools documenting the JSON structure, field names, types, and which fields are always present vs optional. Example for shopee_extract: {"product_id": string, "title": string, "price": {"amount": number, "currency": string}, "seller": {"id": string, "name": string}, "rating": number, "available": boolean, "region": string}
• Rewrite install_mcp_tools description to clarify: (a) is this a one-time server setup step or an on-demand tool? (b) What environment variables or config files must exist? (c) What does success/failure look like? Or, move this logic to server initialization and remove it from the tool list.
• Convert book_to_skill 'format' parameter to an enum: ["epub", "pdf", "docx", "rtf", "technical"] with description 'Document format; determines parser and output structure'
• Add error handling guidance to all scraper tool descriptions: 'Returns error with code 'invalid_url' if URL cannot be parsed; 'unsupported_region' if region not supported; 'timeout' if page takes >30s to load; 'rate_limit' if called >10 times/min. Retry timeouts after 5s; ask user to verify URL on parse errors.'
• For skill_router, add descriptions to 'suggest' and 'categories' boolean params: 'suggest (bool): If true, return skill name suggestions instead of ranked matches. Use when the query is very vague. [true/false; default: false]' and 'categories (bool): If true, also rank skill categories by relevance alongside individual skills. [true/false; default: false]'
• For web scrapers, add compliance warnings in descriptions: '⚠️ Verify you have permission to scrape this platform. Some regions restrict scraping (EU/GDPR, China/WeChat ToS). Respect rate limits (max 10 requests/min) to avoid IP blocks.'