# Adobe Commerce Support MCP Server

This MCP server helps generate professional Adobe Commerce support responses from case findings.

## Installation

1. **Dependencies are already installed** ✅
2. **Server is ready to use** ✅

## Usage with Cursor

To use this MCP server with Cursor, add the following configuration to your Cursor settings:

### Option 1: Add to Cursor Settings (Recommended)

1. Open Cursor Settings (Cmd+,)
2. Search for "MCP" 
3. Add this server configuration:

```json
{
  "mcpServers": {
    "adobe-support-mcp": {
      "command": "python3",
      "args": ["/Users/kavingas/VSCodeProjects/mcp_support_server/mcp_support_server.py"],
      "env": {}
    }
  }
}
```

### Option 2: Use the provided config file

The `mcp_config.json` file in this directory contains the ready-to-use configuration.

## How to Use

### Option 1: Structured Content (Traditional)
1. Create a `find.md` file with properly structured content:

```markdown
### Findings
[Your investigation findings here]

### Analysis  
[Your technical analysis here]

### Recommendations
[Your recommended solutions here]
```

2. Use the `generate_support_reply` tool to create a professional customer response.

### Option 2: Mixed Content (New - LLM Assisted)
1. Create a `find.md` file with mixed/unstructured content (any format)
2. Use either:
   - `categorize_mixed_content` tool first, then use LLM to process the categorization prompt
   - `generate_support_reply` with `auto_categorize=true` (default) for automatic handling

The server will automatically detect if your content needs categorization and provide LLM prompts to organize it properly.

## Tools Available

- **generate_support_reply**: Generates a professional Adobe Commerce support response from case findings
  - `find_file`: Input file (default: "find.md")  
  - `resp_file`: Output file (default: "resp.md")
  - `auto_categorize`: Automatically handle mixed content (default: true)

- **categorize_mixed_content**: Creates LLM prompts to categorize mixed content into structured format
  - `find_file`: Input file with mixed content (default: "find.md")
  - `categorized_file`: Output file with categorization prompt (default: "categorized.md")

## Files

- `mcp_support_server.py` - Main MCP server
- `requirements.txt` - Python dependencies  
- `mcp_config.json` - Cursor configuration
- `setup.py` - Installation script