Cub Query

Use when the user wants to find, count, inspect, read, or audit Kubernetes workloads and application configuration stored in ConfigHub — both fleet-wide sweeps and single-workload lookups. Natural phrasing is workload- or application-centric: "where is checkout v0.4.2 deployed?", "which Deployments run more than 5 replicas?", "find workloads missing resource requests or limits", "what image tag is our nonprod worker running?", "how many replicas does our frontend have in us-east?", "is our api up to date with the latest release?", "show me the YAML / env vars / annotations of our frontend". ConfigHub-native phrasing is equally in scope: "which Units ...", "what''s in Space X", "Units with Label env=prod". Workload-in-environment phrasing like "frontend in us-east" maps to a Unit slug plus a Space slug or Space Label — see space-topology for the conventions. Load any time intent is find / list / show / which / where / how many / audit / inspect / read / what tag / is X up to date over ConfigHub-managed workloads — one workload or the whole fleet. Do not load for: mutating data (cub-mutate), authoring (config-as-data), designing the Space/Label taxonomy itself (space-topology), or live cluster state not in ConfigHub (kubectl).

mkdir -p ~/.claude/skills/cub-query
curl -L https://claudskills.com/skills/cub-query/SKILL.md \
  -o ~/.claude/skills/cub-query/SKILL.md

@misc{cub-query-2026,
  author    = {confighub},
  title     = {Cub Query [Claude Code skill]},
  year      = {2026},
  publisher = {ClaudSkills},
  url       = {https://claudskills.com/skills/cub-query/}
}