WhatsApp Web MCP has complete input schemas with Zod validation and basic descriptions for all 7 tools. However, descriptions are minimal (10-50 chars), lacking context on WHEN to use each tool or dependencies between them. Parameter descriptions exist but are terse. No output schema documentation. Error handling returns generic text without recovery guidance. Tools lack annotations (readOnlyHint/destructiveHint). The send_message tool modifies state but description doesn't explicitly warn of side effects. Resource URIs are returned but not documented in tool output specs.
Loads chat messages, sorted from earliest to latest
Get chat instance by ID
Get all current chat instances
Get contact instance by ID
Get all current contact instances
Searches for messages
Send a message to a specific chatId
Minimal tool descriptions (10-50 chars) lack context on WHEN to use each tool and dependencies. 'Get contact instance by ID' doesn't explain when to call it vs get_contacts, or that it requires a valid contactId or chatId.
No output schema documentation. Tools return resource URIs and JSON-serialized objects, but LLMs cannot infer the structure of returned contact/chat/message objects or which fields are available for downstream tool calls.
send_message modifies state (sends a message) but description does not explicitly state this is a destructive operation. No tool annotations (destructiveHint) present. Error handling returns generic 'Error: {message}' without recovery guidance.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 48 | 2026-07-28+ | v2 |
Pagination parameters (limit, page) present in get_contacts and get_chats but no total count or next_cursor returned. LLMs cannot determine if more results exist or how to iterate safely.
Parameter 'options' in send_message is a nested object with optional fields but no documentation of which combinations are valid (e.g., can you send both media and caption? What if mediaId is invalid?). Undocumented parameter relationships.