stephan48: Not sure what editor you use, but I wanted CRD defintion completions in helix, so I did a thing:
import json
import subprocess
def kubectl_get(path: str) -> dict:
    p = subprocess.run(["kubectl", "get", "--raw", path], capture_output=True)
    return json.loads(p.stdout)
all_types = []
paths = kubectl_get("/openapi/v3")
for name, path in paths['paths'].items():
    print(f"Extracting {name}")
    name = name.replace("/", "-")
    schema = kubectl_get(path['serverRelativeURL'])
    if 'components' in schema and 'schemas' in schema['components']:
        updated_schema = { "components": { "schemas": schema["components"]["schemas"] } }
        for k, crd in schema['components']['schemas'].items():
            if 'type' in crd and crd['type'] == 'object' and 'properties' in crd and 'apiVersion' in crd['properties'] and 'kind' in crd['properties']  and 'x-kubernetes-group-version-kind' in crd:
                kind_enum = []
                api_version_enum = []
                for gvk in crd["x-kubernetes-group-version-kind"]:
                    kind_enum.append(gvk['kind'])
                    if gvk['group'] == '':
                        api_version_enum.append(gvk['version'])
                    else:
                        api_version_enum.append(f"{gvk["group"]}/{gvk["version"]}")
                crd["properties"]["kind"]["enum"] = kind_enum
                crd["properties"]["apiVersion"]["enum"] = api_version_enum
                all_types.append({ "$ref": f"{name}.json#/components/schemas/{k}" })
            updated_schema['components']['schemas'][k] = crd
        with open(f".k8s-schema/{name}.json", "w") as f:
            f.write(json.dumps(updated_schema))
with open(".k8s-schema/all.json", "w") as f:
    f.write(json.dumps({ "oneOf": all_types }))