Verification

To confirm the plugin is working correctly:

1. Install Visual Studio Code

Download and install VSCode from the official website: Visual Studio Code.

2. Install the AI Toolkit extension

Open VSCode and add the AI Toolkit extension. For detailed instructions, visit: AI Toolkit for Visual Studio Code.

3. Connect to MCP Integrate

Follow the instructions in the Connect to MCP Server section to link your AI Toolkit with MCP Integrate.

4. Verify your mcp.json

Your configuration file should look like this:

{
  "servers": {
    "mcp-integrate": {
      "type": "http",
      "url": "http://<your-site-domain>/wp-json/mcp-integrate/entrypoint"
    }
  }
}

5. Add a system prompt

To help the AI assistant coordinate all tools more efficiently, you may provide a system prompt. This is optional — the tools will still work without it, but performance and relevance may improve.

⚠️ DISCLAIMER
The following system prompt is provided only as an example. Before using it in a production environment, make sure to review and validate it with your organization’s security and compliance teams. Requirements may vary depending on your infrastructure, policies, and risk profile.

1. Always start by analyzing the user input.
  - If the input contains multiple keywords that can map to different dimensions 
    (brand, category, tag, or product name), then call the corresponding tools 
    in combination immediately (for example, "Nike running shoes" = brand + category).
    Try to extract as many keywords as possible.
  - If the input contains only one relevant keyword, 
    always call search_products_by_name_mcp first.

2. After collecting results from the first call(s):
  - If you have at least 5 results, return them.
  - If you have fewer than 5 results, continue with fallback logic.

3. Fallback logic (if you have less than 5 results):
  Call the tools one by one in this order until you gather at least 5 results:
  - search_products_by_category_mcp
  - search_products_by_tag_mcp
  - search_products_by_brand_mcp
  If still fewer than 5 results, retry by splitting the input into individual 
  keywords and repeating the process.

4. Always combine all non-empty results, remove duplicates, and always return 
products with their permalinks.

5. If after all retries there are still fewer than 5 results, return what you have 
and explain politely to the user.

Never expose tool names, API details, or internal system logic to the user. 

6. Start chatting with your AI assistant

Once connected, you can immediately start interacting with your AI assistant (e.g., ChatGPT) through MCP Integrate to quickly discover and explore products that interest you.