Support to define multiple platforms

when initializing a new scenario for the azure driver a create.yml is created inside the newly created scenario

when looking at the create.yml the task to create the molecule instance has a CentOs hard-coded instance into the task, my request is: can the template that creates this create.yml be updated to take all the values from the molecule_yml.platforms?, that way we don't have to manually update the create.yml and just provide all the platforms specifics in the platforms section of the molecule.yml

here is the task for the Create molecule instance(s) (taken from the auto-created create.yml)

```yaml
- name: Create molecule instance(s)
azure_rm_virtualmachine:
resource_group: "{{ resource_group_name }}"
name: "{{ item.name }}"
vm_size: Standard_A0
admin_username: "{{ ssh_user }}"
public_ip_allocation_method: Dynamic
ssh_password_enabled: false
ssh_public_keys:
- path: "/home/{{ ssh_user }}/.ssh/authorized_keys"
key_data: "{{ key_pair.ssh_public_key }}"
image:
offer: CentOS
publisher: OpenLogic
sku: '7.4'
version: latest
register: server
with_items: "{{ molecule_yml.platforms }}"
async: 7200
poll: 0
```

Start by locating the template that generates the Azure scenario's create.yml and compare it with the platforms section in molecule.yml. Trace how molecule_yml.platforms is passed to the create task, then verify that each platform's image settings are used instead of the hard-coded CentOS values.

Written by the indexing model from the issue text.