Server has 5 tools with bilingual (Mongolian/English) descriptions and basic parameter schemas. All tools are READ_ONLY and well-intentioned for a classifieds scraper. However, descriptions are inconsistent in depth, parameter descriptions lack constraint details (enums, ranges, formats), output schemas are not formally documented, and error handling is minimal. Tool names follow verb_noun convention (search_, browse_, get_, list_) which is good. Parameter defaults are present but some lack clarity on valid values. No tool annotations (readOnlyHint, etc.) despite all being read-only. Average tool score: 62/100.

• Add formal output schema documentation for each tool. Define the JSON structure returned (e.g., {status, query, category, page, count, listings: [{title, price, url, location, date}]}). This enables LLMs to extract fields correctly and chain tools.
• Convert 'category' parameter to an enum constraint with values: vehicles, real_estate, electronics, jobs, services, clothing, furniture, pets, hobby, education. Remove examples from description and move to schema.
• Add min/max bounds to numeric parameters: page (min: 1), limit (min: 1, max: 50). Include these in parameter descriptions and schema.
• Enhance error messages with recovery guidance. E.g., 'Category "invalid_cat" not found. Valid categories: vehicles, real_estate, electronics, jobs, services, clothing, furniture, pets, hobby, education.' or 'No listings found. Try a different search term or browse_category() to explore available items.'
• Add readOnlyHint: true annotation to all 5 tools in the tool definition to signal they are safe for caching and retry.