domdomegg

Call MCP — AI skill for Claude Code

AI community

📞 CLI to call your stdio, shttp and claude.ai-configured MCP servers.

How to install Call MCP

This entry records only its repository, not the path inside it, so there is no exact command to give. Open domdomegg/call-mcp and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Call MCP does

📞 CLI to call your stdio, shttp and claude.ai-configured MCP servers.

Alternatives in AI

MCP Use — The fullstack MCP framework to develop MCP Apps for ChatGPT / Claude & MCP Servers for AI Agents 9.5k ★
Rulesync — by dyoshikawa - A Node.js CLI tool that automatically generates configs (rules, ignore files, MCP servers, com 909 ★
Codeseek — Rust-powered code intelligence CLI for AI coding agents 764 ★

README

call-mcp

A CLI for calling MCP servers: define servers in a small config file (MCP Streamable HTTP or stdio, including OAuth-secured servers), and/or use the connectors you've set up in claude.ai (those reuse the Claude Code OAuth token — no separate login).

Every command prints JSON to stdout, on success and on error alike, so output is always safe to pipe into `jq` or parse in a script.

Install

npm install -g call-mcp

Requires Node 20+. Works on macOS, Linux, and Windows. Servers from your own config need no other setup; claude.ai connectors need a Claude Code login (or `CLAUDE_CODE_OAUTH_TOKEN`).

Usage

call-mcp list                                List your servers and claude.ai connectors
call-mcp tools                       List the tools a server exposes
call-mcp call   [--args JSON]  Call a tool
call-mcp help                                Full help with examples

`` is a server name from your config file, a claude.ai connector's display name, or a connector id (`mcpsrv_...`) — a unique case-insensitive prefix also works, e.g. `google`.

`--args ` passes a JSON object of arguments to `call`.

`--full` includes full detail in the output. By default call-mcp prints slim objects; `--full` adds tool input/output schemas, raw server fields, and the complete tool-call result envelope. Keep it slim and search first — fetch a full schema only for the tool you actually need.

Configuring servers

Define your servers in `~/.config/call-mcp/servers.json` (or point `CALL_MCP_SERVERS_FILE` at a file), using the same `mcpServers` shape as Claude Code's MCP config — blocks copy-paste between the two:

{
  "mcpServers": {
    "homelab": {
      "type": "http",
      "url": "https://mcp.example.com/mcp",
      "headers": { "Authorization": "Bearer ${MY_TOKEN}" }
    },
    "everything": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-everything"]