# Method: method-cluster-cr

Translate curated Cluster evidence intents into toolkit `get-cluster-cr`.

## When to use

- Evidence id `cluster_cr` or `scale_status_cm`
- Procedure needs Cluster CR by `cluster_id` (not arbitrary Kind Get)

## Bind / execute

1. Read toolkit subskill:
   `skills/utilities/dedicated-cloud-toolkit/get-cluster-cr/SUBSKILL.md`
   (and its prerequisite mechanism as that file requires).
2. Call Get Cluster CR with `cluster_id` (+ optional `env`).
3. Format via that subskill's `scripts/format-cluster-resources.sh`.

Do **not** re-document full API fields here. Follow the toolkit SUBSKILL.

## Condition → resources

| Intent condition | Resource key | Notes |
|------------------|--------------|-------|
| by `cluster_id` (default) | `cluster_cr` | Always expected when API succeeds |
| scale status needed | `scale_status_cm` | Present only when scale-status feature enabled; absence may be soft or case-specific signal |

## Diagnosis focus (after fetch)

From `cluster_cr` content, prefer:

- `status.phase`
- `spec.networkName` (→ APPBox / network ref via object-graph)
- `spec.kubernetes.provider` (→ provider context)
- conditions / messages if present
- component replica / instance fields relevant to the case

From `scale_status_cm` (when present): scale phase / progress fields named in
the case Procedure.

## Shape choice

This method is always curated-by-`cluster_id`. Prefer it over
`method-k8s-resource` for Cluster CR / scale CM. For APPBox, Pods, TiDBCluster,
or label lists, use `method-k8s-resource`.