How does the approval process work?
When Claude attempts to execute a tool, the `permissions__approve` function sends a WhatsApp message with formatted command details and quick‑reply buttons. Your tap approves or denies execution.
@adstastic
基本信息
MCP server with Whatsapp integration (Twilio) to remote approve Claude Code tool use
使用下面的配置,将此服务器添加到你的 MCP 客户端。
{
"mcpServers": {
"claude-code-whatsapp-approval": {
"command": "python",
"args": [
"setup_template.py"
]
}
}
}工具列表暂未提供
当前收录暂时没有可展示的工具列表,不代表服务没有工具。请查阅项目文档了解支持的能力。
工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。
A Model Context Protocol (MCP) server that provides approval prompts via WhatsApp before executing potentially sensitive commands. Built with FastMCP and Twilio, it is designed for users of Claude who want a security layer on tool execution.
Install dependencies with uv install, set up a Twilio account, configure environment variables in .env, create a WhatsApp template, expose your local server via ngrok or Tailscale, configure the Twilio webhook, start the server with uv run approval_server.py, then launch Claude with --mcp-config mcp-servers.json and --permission-prompt-tool mcp__approval-server__permissions__approve. Approve or deny commands by tapping WhatsApp quick-reply buttons.
When Claude attempts to execute a tool, the `permissions__approve` function sends a WhatsApp message with formatted command details and quick‑reply buttons. Your tap approves or denies execution.
The request expires after 5 minutes for security. If no response is received, Claude does not execute the command.
No. Non‑interactive mode (without the `-p` flag) does not use the `--permission-prompt-tool` flag, so approval prompts are not triggered in that mode.
The phone number set in the `APPROVAL_PHONE` environment variable. This must be the same number that joined the Twilio WhatsApp sandbox.
The `-p` (prompt) mode itself has no network access, but the MCP server runs as a separate process and can still send WhatsApp requests if it has network connectivity.

1,000+ fully localized prompt templates across 40+ categories. Compare Mode — send the same prompt to up to 9 AI providers (ChatGPT, Claude, Gemini, Grok, Groq, Copilot, Mistral, Perplexity, DuckDuckGo AI) and compare an

Docs: https://botchelor.com/agents and https://botchelor.com/llms.txt . Botchelor is a blind-date show for fictional characters, and AI agents are welcome to play. Over MCP, your agent builds an original character in w

Hire AI agents and lock USDC in trustless on-chain escrow on Solana all from your AI assistant. Browse agents by skill or protocol, prepare a hire, and track escrow status without leaving your chat.
NetFoundry MCP Gateway enables secure, isolated access to Model Context Protocol (MCP) tools across distributed systems without exposing public endpoints. Zero trust gateway for MCP servers. Aggregate, filter, and secure

scvd.store is an evidence observatory for agentic commerce. Independent signed observation of what other people's endpoints, artifacts and payments actually did among other things. USDC on Base, Arbitrum One, World Chain
评论