cursor_integration.md•2.96 kB
# Cursor Integration Guide for kubectl-mcp-tool
This guide explains how to integrate the kubectl-mcp-tool with Cursor AI assistant for natural language Kubernetes operations.
## Prerequisites
- kubectl-mcp-tool installed
- Cursor AI assistant
- Python 3.8+
## Setup
1. Install the kubectl-mcp-tool:
```bash
pip install -e /path/to/kubectl-mcp-tool
```
2. Start the MCP server with Cursor compatibility mode:
```bash
python -m kubectl_mcp_tool.cli serve --cursor
```
3. Configure Cursor to use the kubectl-mcp-tool:
- Open Cursor AI assistant
- Go to Settings > Tools
- Add a new tool with the following configuration:
- Tool Name: `kubectl-mcp-tool`
- Command: `python -m kubectl_mcp_tool.cli serve --cursor`
- Working Directory: `/path/to/kubectl-mcp-tool`
## Available Tools
The kubectl-mcp-tool provides the following tools through the MCP interface:
| Tool Name | Description | Parameters |
|-----------|-------------|------------|
| `process_natural_language` | Process natural language queries for kubectl operations | `query`: The natural language query to process |
| `get_pods` | Get all pods in a namespace | `namespace`: The namespace to get pods from (optional) |
| `get_namespaces` | Get all namespaces in the cluster | None |
| `switch_namespace` | Switch to a different namespace | `namespace`: The namespace to switch to |
| `get_current_namespace` | Get the current namespace | None |
| `get_deployments` | Get all deployments in a namespace | `namespace`: The namespace to get deployments from (optional) |
## Example Natural Language Commands
You can use the following natural language commands with Cursor:
- "Get all pods"
- "Show namespaces"
- "Switch to namespace kube-system"
- "What is my current namespace"
- "Get deployments"
- "Get services"
- "Describe pod nginx-pod"
## Troubleshooting
If you encounter issues with the Cursor integration, try the following:
1. **Client Closed Status**: If Cursor shows "Client closed" status, try restarting Cursor and the MCP server.
2. **Check Logs**: Look for error messages in the following log files:
- `simple_mcp_server.log`
- `simple_mcp_debug.log`
- `kubectl_mcp_tool_cli.log`
3. **Test the Server**: Run the test script to verify the MCP server is working correctly:
```bash
python test_simple_mcp_server.py
```
4. **Cursor Configuration**: Ensure the tool is registered correctly in Cursor's configuration.
5. **Direct Server Run**: Try running the simple MCP server directly:
```bash
python simple_mcp_server.py
```
## Advanced Configuration
For advanced configuration options, see the [Configuration Guide](./configuration.md).
## Example Usage in Cursor
Once configured, you can use the kubectl-mcp-tool in Cursor by typing natural language commands like:
```
Get all pods in the default namespace
```
Cursor will use the kubectl-mcp-tool to execute the command and return the results.