WhatsApp MCP has 25 tools with consistent verb-noun naming (search_, list_, get_, send_, create_, etc.). All tools have descriptions (avg ~80 chars, within baseline 34-392 range). Input schemas are present and typed for all tools. However, descriptions are often terse and lack context about when to use each tool vs. alternatives. Output schemas are not documented, LLMs cannot predict response structure. Error handling is absent from descriptions. No tool annotations (readOnlyHint/destructiveHint) despite clear risk stratification (READ_ONLY, WRITE, DESTRUCTIVE). Parameter descriptions are minimal; many lack format/constraint guidance. No pagination guidance in list_* tools despite accepting limit/page params. Missing idempotency guarantees for write operations.

• Add output schemas to all tools. Document the structure of responses (e.g., send_message returns {message_id: string, timestamp: ISO8601, status: 'sent'|'pending'|'failed'}).
• Add tool annotations to schema: mark all WRITE tools with 'destructiveHint: false' and DESTRUCTIVE tools with 'destructiveHint: true'. Mark READ_ONLY tools with 'readOnlyHint: true'. This enables agents to reason about retry safety.
• Expand parameter descriptions with format constraints. Replace 'phone number with country code but no + or other symbols' with an enum or regex pattern in the JSON schema, and reference it in the description.
• Document pagination behavior: 'Returns up to {limit} results (default 20, max 100). Use page param to iterate. Response includes total_count for planning.'
• Add error recovery hints to tool descriptions. E.g., send_message: 'If recipient not found, call search_contacts() first. If message fails, check network and retry. Max message length is 4096 characters.'

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

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