gke-pod-restarts

name	gke-pod-restarts
description	Use this skill to investigate GKE incidents involving pod restart cascades, 503/504 errors, service unavailability, or CrashLoopBackOff. Triggers: 'pods keep restarting', '503/504 errors', 'service is down', 'crash loop', 'what caused the incident at <time>'. Inputs required: GCP project, cluster name+region, namespace, service/deployment names, approximate incident time + timezone. All commands are read-only. Never mutate cluster state.
model	gpt-5.4
effort	high

Read ~/.ai-shared/skills/gke/pod-restarts.md and follow all instructions exactly.