Docs: Add a guide for using kubectl-ai in a production environment

We have been getting a few requests around best practices around using kubectl-ai safely in production environment.

Task here is to create a simple guide sharing tips/tricks for enabling safe usage of kubectl-ai in production.

Some quick notes:

• Run in a least privilege environment
• Running in sandbox
• Using LLM provider such as vertexai with enterprise controls around data privacy etc.
• tips on command approvals