FXMacroData server has 8 well-named tools with consistent verb_noun patterns (ping, data_catalogue, release_calendar, forex, indicator_query, market_sessions, cot_data, commodities). All tools have descriptions (avg ~120 chars, within 10-1024 range). However, input schemas are incomplete: most parameters lack explicit type definitions in the visible code (only inferred from function signatures). Parameter descriptions are present but minimal (10-30 chars). Output schemas are not documented. Error handling exists but is generic (JSON error wrapper without recovery guidance). No tool annotations (readOnlyHint, etc.). Composition is sound, each tool has one responsibility, and naming enables chaining (e.g., data_catalogue → indicator_query). Security: API key injected via environment variable (good), but no rate limiting or input validation visible.

• Add explicit JSON Schema type definitions to all tool inputs. Use FastMCP's schema parameter or manually construct JSON Schema objects with 'type', 'description', 'enum', 'pattern', and 'minLength'/'maxLength' for each parameter.
• Document output schemas for each tool. Add a 'Returns' section to each tool's docstring specifying the structure of the JSON response (e.g., 'Returns: {"indicators": [{"slug": string, "name": string, ...}], "currency": string}').
• Enhance error handling to classify errors. Return structured errors with 'error_type' (e.g., 'invalid_input', 'api_error', 'auth_required', 'not_found') and 'recovery_hint' (e.g., 'Try data_catalogue(currency) first to discover valid indicators').
• Add tool annotations. Declare all tools with readOnlyHint=true since they are read-only. This enables clients to optimize caching and concurrency.
• Expand parameter descriptions to 50-200 chars. For 'indicators' in forex, list valid slugs: 'Comma-separated technical indicators: sma_20, ema_50, rsi_14, bbands_20, atr_14, macd. E.g., "sma_20,rsi_14"'. For 'currency', add format: 'ISO 4217 currency code (3 uppercase letters, e.g., USD, EUR, GBP)'.