WhatsApp MCP has 15 tools with basic schemas and descriptions, but significant gaps prevent higher scores. All tools have descriptions (10-50 chars, mostly short), and input schemas are present with type definitions. However, descriptions lack LLM-optimized guidance on WHEN to use each tool vs. alternatives, parameter relationships are undocumented (e.g., mediaUrl vs mediaPath in send_message), and output schemas are entirely absent from the source. Error handling is minimal, no recovery guidance, no actionable error messages, no categorization of retryable vs. fatal failures. No tool annotations (readOnlyHint/destructiveHint) despite clear risk levels. Naming is generally good (verb_noun pattern), but some tools conflate concerns (whatsapp_setup handles 7 distinct actions via enum). The server lacks pagination support despite list_chats accepting a limit parameter with no documented max or default.

• Document output schemas for all tools. Specify field names, types, and whether arrays are paginated. E.g., whatsapp_list_chats should return {chats: [{id, name, unreadCount, lastMessage}], total: number}.
• Expand descriptions to 50-200 chars with WHAT/WHEN/WHY guidance. E.g., 'Get detailed connection status including auth state, battery, and network. Call this first to verify WhatsApp is ready before sending messages.'
• Add explicit parameter constraints in descriptions. E.g., 'mediaUrl and mediaPath are mutually exclusive, provide one or the other, not both. mediaUrl fetches from URL; mediaPath uses local file.'
• Implement error categorization: return {error: string, category: 'retryable'|'user_fixable'|'fatal', suggestion: string}. E.g., 'Chat not found. Try search_chats() with a partial name.' for user_fixable.
• Split whatsapp_setup into separate tools: list_profiles, add_profile, remove_profile, switch_profile, clone_session. Each should have a single, clear responsibility.

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

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