Gmail MCP server has 7 tools with basic schemas and descriptions in Spanish. All tools have input schemas with type definitions and descriptions present. However, descriptions are minimal (10-50 chars), lack context on WHEN to use each tool, and do not explain consequences of write operations. No output schemas documented. Parameters lack constraints (enums, ranges, formats). Error handling and recovery guidance absent. Tool composition is reasonable (separate read/write operations), but descriptions do not meet LLM-optimization baselines (target 50-200 chars). No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk stratification.

• Expand all descriptions to 50-200 chars. Example: 'send_email' → 'Send a new email to one or more recipients. Use this to initiate new conversations. For replies, use reply_to_email. Irreversible, confirm before sending to prevent accidental messages.'
• Document output schemas for all tools. Example: list_emails returns {messages: [{id, from, subject, date, snippet}], total_count, next_cursor}. This enables agents to chain calls (e.g., get_email after list_emails).
• Add tool annotations: send_email, reply_to_email, create_draft → destructiveHint=true (irreversible); list_emails, get_email, search_emails → readOnlyHint=true; delete_email → destructiveHint=true, idempotentHint=false (reversible but not idempotent).
• Add parameter constraints: maxResults type=integer, minimum=1, maximum=50; labelIds type=array, items={type: string}, maxItems=10; query type=string, maxLength=1000.
• Implement error handling: return structured errors with category (retryable|user_fixable|fatal), invalid_value, constraint_violated, and recovery_hint. Example: {error: 'Invalid recipient', category: 'user_fixable', hint: 'Verify email address format ([email protected])'}.

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

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