SerpBase MCP exposes 6 search tools with consistent schemas and basic descriptions. All tools have input schemas with proper JSON Schema structure (type, properties, required fields). Descriptions are present but generic, they state WHAT the tool does but lack WHEN to use it, dependencies, or output structure guidance. Parameter descriptions are minimal (e.g., 'Search query' for query param). No output schemas documented. Error handling is basic (generic exception wrapping). Tools follow a repetitive pattern (query + hl/gl/page), which aids consistency but limits expressiveness. No tool annotations (readOnlyHint, destructiveHint) despite all being read-only operations. Naming is clear and verb-prefixed (serpbase_*), but descriptions could be 2-3x longer to guide LLM selection.

• Expand tool descriptions to 150-250 chars. Include WHEN to use (e.g., 'Use for organic web results; use serpbase_images for visual search'). Add output hints (e.g., 'Returns organic results array with title, link, snippet, and rich SERP modules like People Also Ask').
• Document parameter constraints in descriptions. E.g., 'hl: ISO 639-1 language code (e.g., en, es, fr); defaults to en' and 'page: 1-indexed; max 100 pages per API limits'.
• Add readOnlyHint=true to all 6 tools in types.Tool() definitions. This signals to agents that these calls are safe to retry and have no side effects.
• Create a separate tool or extend responses to document output schema. E.g., add a 'schema' field to tool definitions or publish a separate schema reference. At minimum, document in descriptions: 'Returns {results: [{title, link, snippet, ...}], total_results: number, ...}'.
• Improve error handling in call_tool(). Catch SerpBaseError separately and return structured errors with recovery hints: {'error': 'API key not set', 'recovery': 'Set SERPBASE_API_KEY environment variable', 'docs': 'https://serpbase.dev/...'} instead of raw exception strings.