Futures Calc MCP demonstrates solid fundamentals: all 4 tools have clear verb-noun naming (get_, calculate_), descriptions between 80-180 chars, and complete Zod schemas with typed parameters. Input schemas are well-structured with descriptions for every parameter. However, output schemas are not formally documented, responses are free-text markdown strings rather than structured objects. Error handling is minimal (only unknown symbol checks). No tool annotations (readOnlyHint, idempotentHint) despite all tools being read-only. Parameter descriptions lack format constraints (e.g., 'risk_pct' should specify range 0-100). Missing output documentation forces LLMs to parse unstructured text.

• Convert all tool responses from markdown text to structured JSON objects. E.g., get_contract_specs should return {symbol, name, tickSize, tickValue, pointValue, margin, currency} instead of formatted text.
• Add numeric constraints to parameter descriptions: 'risk_pct: number (0-100, e.g., 1 for 1%)', 'stop_ticks: positive integer', 'contracts: positive integer (default 1)'.
• Add readOnlyHint: true and idempotentHint: true to all tool definitions via Zod metadata or MCP tool registration options.
• Enhance error handling: validate numeric ranges before calculation, return structured error objects with error_code and recovery_hint fields, e.g., {error: 'invalid_risk_pct', message: 'risk_pct must be 0-100', recovery: 'Provide a percentage between 0 and 100'}.
• Document output schema for each tool in the tool description or via a separate schema field. E.g., 'Returns: {symbol: string, name: string, tickSize: number, tickValue: number, pointValue: number, margin: number, currency: string}'.