JustSerpAPI MCP has well-structured tool definitions with clear naming (verb_noun pattern) and comprehensive descriptions (avg 150+ chars). All 7 tools have input schemas with type definitions. However, parameter descriptions are sparse or missing in several tools (e.g., call_endpoint's 'params' object lacks field-level documentation), and output schemas are not documented. Error handling exists but lacks recovery guidance. The server follows a logical discovery workflow (search → schema → call) but does not fully leverage structured output patterns or tool annotations.

• Document output schemas for all tools. For search_endpoints, specify: { endpoint_id: string, confidence: number, service: string }. For get_account_balance: { subscription_tier: string, remaining_credits: number, total_credits: number }. This enables LLMs to chain calls without guessing.
• Enhance call_endpoint's 'params' parameter with a description that explains: 'Object with snake_case keys matching the endpoint schema. Required fields vary by endpoint_id; call get_endpoint_schema first to see the full contract.' Consider adding an example or link to schema documentation.
• Improve error messages in runTool() to include recovery hints. E.g., 'API key invalid or expired. Check JUSTSERPAPI_API_KEY environment variable.' or 'Endpoint not found. Call search_endpoints to discover available endpoints.'
• Add tool annotations (readOnlyHint, destructiveHint) to tool definitions. Mark search_endpoints, get_endpoint_schema, get_account_balance, get_usage_summary, list_services as readOnly=true. Mark refresh_catalog as destructive=true (modifies catalog state).
• Implement explicit permission validation in refresh_catalog. If isAdmin() returns false, return a structured error: { code: 'PERMISSION_DENIED', message: 'refresh_catalog requires admin privileges' } instead of silently failing or succeeding.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry