Israeli MCP Servers Guide

What is MCP?

Model Context Protocol (MCP) is an open standard created by Anthropic that lets AI agents connect to external data sources and tools. Instead of copying and pasting data into the chat, an MCP server exposes APIs as structured tools that the agent can call directly.

For example, an MCP server for weather data lets the agent pull a live forecast from IMS (Israel Meteorological Service) without you having to browse the website yourself.

The Skills IL MCP Directory

Our MCP Server Directory is a curated collection of open-source MCP servers that connect AI agents to Israeli data. Servers are organized by category:

Category	Examples
Government & Public Data	Knesset, Tax Authority, data.gov.il
Finance & Banking	Bank of Israel, Stock Exchange, exchange rates
Statistics & Economics	CBS (Central Bureau of Statistics), macro data
Healthcare & Medical	Ministry of Health, HMOs
Real Estate & Land	Tabu (Land Registry), Israel Land Authority
Transportation	Israel Railways, buses, traffic
Weather & Environment	IMS, air quality
Library & Archives	National Library

Each server is marked by status: Available (working and tested) or Community (built by the community).

Installation: Step by Step

Step 1: Choose a Server

Browse the MCP Server Directory and find a server that fits your needs. Click on a server card to see full details.

Step 2: Configure Your Agent

Each MCP server requires a configuration entry in your agent's settings file. The server page shows the exact configuration for each supported agent.

Claude Code (.mcp.json in your project folder):

{
  "mcpServers": {
    "israel-weather": {
      "command": "npx",
      "args": ["@skills-il/mcp-weather"]
    }
  }
}

Cursor (.cursor/mcp.json):

{
  "mcpServers": {
    "israel-weather": {
      "command": "npx",
      "args": ["@skills-il/mcp-weather"]
    }
  }
}

Windsurf (~/.codeium/windsurf/mcp_config.json):

{
  "mcpServers": {
    "israel-weather": {
      "command": "npx",
      "args": ["@skills-il/mcp-weather"]
    }
  }
}

Step 3: Environment Variables

Some servers require API keys or environment variables. This information appears on the server page and in its GitHub documentation. Example:

{
  "mcpServers": {
    "bank-of-israel": {
      "command": "npx",
      "args": ["@skills-il/mcp-boi"],
      "env": {
        "BOI_API_KEY": "your-api-key-here"
      }
    }
  }
}

Step 4: Verify

After adding the configuration and any required environment variables, restart your agent. You can verify the server is connected by asking your agent:

"What MCP tools are available to you?"

The agent should list the tools the server exposes.

Using Multiple Servers Together

MCP is designed to be composable. You can configure multiple servers together, and the agent gets access to all their tools:

{
  "mcpServers": {
    "israel-weather": {
      "command": "npx",
      "args": ["@skills-il/mcp-weather"]
    },
    "knesset": {
      "command": "npx",
      "args": ["@skills-il/mcp-knesset"]
    },
    "cbs": {
      "command": "npx",
      "args": ["@skills-il/mcp-cbs"]
    }
  }
}

This way the agent can, for example, pull statistical data from CBS, cross-reference it with Knesset legislation, and check the weather forecast, all in a single conversation.

Debugging with MCP Inspector

MCP Inspector is a visual tool for testing and debugging MCP servers. Use it when something is not working as expected.

Installation and Usage

npx @modelcontextprotocol/inspector

This opens a browser interface that lets you:

• Connect to a server - enter the server's start command and see if it starts successfully
• View tools - list all tools the server exposes, including their parameters
• Run tools - manually call tools with parameters you choose and see the result
• Inspect resources - view resources the server provides
• Monitor messages - watch the client-server communication in real time

Common Scenarios

Server does not connect:

1. Run the Inspector with the same command from your JSON config
2. Check that all environment variables are set
3. Look for error messages in the Inspector log

Tool returns an error:

1. Open the Inspector and call the tool directly
2. Check the parameters you are sending
3. Compare the result with the tool's documentation

Agent does not see the tools:

1. Make sure the config file is in the correct path
2. Restart the agent after updating the config
3. Check for JSON syntax errors in the config file

Tips for Effective Use

1. Start with one server - don't configure all servers at once. Start with one, make sure it works, then add more.

2. Read the documentation - each server offers different tools. Understanding the server's capabilities enables more precise usage.

3. Use clear names - when configuring multiple servers, choose clear key names (israel-weather, knesset, boi) so they are easy to identify.

4. Update regularly - MCP servers receive updates. Check the package changelog on npm or GitHub periodically, and clear your npx cache if you want to force the latest version.

5. Test with Inspector - before reporting a bug, check with MCP Inspector whether the issue is in the server or in your configuration.

Building Your Own MCP Server

You can build an MCP server using the official SDK:

• TypeScript: @modelcontextprotocol/sdk package
• Python: mcp package

For guides and examples, see the official MCP documentation.

If you built an MCP server for Israeli data and want to add it to our directory, reach out to us and we will be happy to include it.