Whatsapp Mcp — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Whatsapp Mcp (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Check it out on https://mcp.paperknife.app/
<!-- mcp-name: io.github.kalki-kgp/whatsapp-macos -->
A Model Context Protocol server that connects Claude to your WhatsApp. Ask Claude to catch you up on unread chats, search your message history, or send a reply.
Reading works in about a minute with no QR scan and no Node.js. Sending is optional and needs one extra step.
<p align="center"> <img src="https://img.shields.io/badge/platform-macOS-blue" alt="macOS"> <img src="https://img.shields.io/badge/MCP-1.0-green" alt="MCP 1.0"> <img src="https://img.shields.io/badge/license-MIT-lightgrey" alt="MIT License"> </p>
Search contacts by name or phone number. Read chat history with date filtering. List recent conversations with unread counts. Summarize everything you haven't read yet. Send replies and receive incoming messages once the bridge is connected.
You don't need to install Python. uvx fetches a suitable version on its own.
Install uv if you don't have it:
curl -LsSf https://astral.sh/uv/install.sh | shThat's the whole install. uvx downloads the server on first launch, so there's no virtualenv to create and no pip install step. Go to the next section to point Claude at it.
git clone https://github.com/kalki-kgp/whatsapp-mcp.git
cd whatsapp-mcp
uv venv && uv pip install -e . open ~/Library/Application\ Support/Claude/claude_desktop_config.jsonIf it doesn't exist, create it.
uvx: which uvx {
"mcpServers": {
"whatsapp": {
"command": "/Users/YOU/.local/bin/uvx",
"args": ["whatsapp-mcp-macos"]
}
}
}Use the absolute path, not bare uvx. Claude Desktop starts with a minimal PATH that doesn't include ~/.local/bin, so a bare command name is the most common reason the server never appears.
Add to .cursor/mcp.json in your project:
{
"mcpServers": {
"whatsapp": {
"command": "/Users/YOU/.local/bin/uvx",
"args": ["whatsapp-mcp-macos"]
}
}
}Restart Cursor and use WhatsApp tools in the AI chat.
Just ask Claude:
cd bridge && npm install && npm start| Tool | Description | Requires Bridge |
|---|---|---|
whatsapp_status | Check connection, get QR if needed | No |
whatsapp_search_contacts | Search contacts by name/phone | No |
whatsapp_list_chats | List recent conversations | No |
whatsapp_get_messages | Get messages from a chat | No |
whatsapp_search_messages | Search across all chats | No |
whatsapp_unread | Get unread message summary | No |
whatsapp_send | Send a message | Yes |
whatsapp_incoming | Get real-time incoming messages | Yes |
Claude ──MCP──▶ WhatsApp MCP Server
│
├──▶ Local SQLite DBs (read messages)
│ ~/Library/Group Containers/group.net.whatsapp.WhatsApp.shared/
│
└──▶ WhatsApp Bridge (:3010) ──▶ WhatsApp Web
(for sending)Read operations query the local WhatsApp database directly, so they work offline and need no bridge.
Send operations go through the bridge, which connects to WhatsApp Web using Baileys.
# Clone
git clone https://github.com/kalki-kgp/whatsapp-mcp.git
cd whatsapp-mcp
# Install in dev mode
pip install -e ".[dev]"
# Run server
python -m whatsapp_mcpMIT
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.