Arlon CLI VS Arlon Service

Right now Arlon CLI and the corresponding Arlon library assumes that its running as a desktop/workstation client and a number of elements exist, including Kubectl, GitHub cli, argocd cli. This approach works for a workstation tool, however introduces issues if we want to run the Arlon Library as a service in the management cluster. The library dependancies also mean the users must have these dependancies installed and configured.

• If Arlon Library runs as a service it will need to write to GitHub, gitlab. How will we handle secrets for auth? how will this be managed on a user by user basis, how do we create an audit trail.