[FEAT cloudctl] - Simplify sync command with persisted Greenhouse connection config #79

User Story

As a cloudctl user, I want to run cloudctl sync without flags after a one-time setup, so that syncing my kubeconfig is a single command rather than a long invocation.

Description

The current sync command requires at minimum --greenhouse-cluster-namespace on every invocation, and in practice users also need --greenhouse-cluster-context and sometimes --greenhouse-cluster-kubeconfig when the Greenhouse cluster is just one context among many in their default kubeconfig. This makes the command too verbose for daily use.

Current typical real-world invocation:

cloudctl sync \
  --greenhouse-cluster-namespace my-org \
  --greenhouse-cluster-context greenhouse \
  --greenhouse-cluster-kubeconfig ~/.kube/config \
  --auth-type exec-plugin \
  --kubelogin-path /path/to/kubelogin

These values are stable per user and should only need to be provided once. The fix is to persist them via cloudctl config (the profile mechanism from the set/env issue) so that cloudctl sync with no flags works after a one-time setup. Additionally, --kubelogin-path should default to the version-pinned kubelogin binary managed by cloudctl (once the logon issue lands) rather than requiring kubelogin to be on $PATH.

Acceptance Criteria

• --greenhouse-cluster-namespace, --greenhouse-cluster-context, --greenhouse-cluster-kubeconfig can be persisted to cloudctl config and used as defaults by sync
• Explicit flags always override persisted config
• --kubelogin-path defaults to the cloudctl-managed version-pinned kubelogin binary when available, falling back to kubelogin on $PATH
• cloudctl sync with no flags and a valid persisted config produces the same result as the full flag invocation
• Missing required config (namespace not set anywhere) produces a clear error pointing to the one-time setup step

Reference Issues

Epic: #70