Diagnose NCCL failures and adjacent training-pod failures on HyperPod GPU clusters (EKS or Slurm) — training hangs, AllReduce / collective-op timeouts, EFA or libfabric errors, rendezvous failures, EFA TCP fallback, /dev/shm or memlock issues, NCCL version mismatch across pods, container OOM / exit-137 / OOMKilled, GPU OOM (CUDA out of memory), CrashLoopBackOff / Pending pods, MASTER_ADDR DNS…
Install
npx skills add awslabs/agent-plugins --skill hyperpod-ncclnpx skills add awslabs/agent-plugins --skill hyperpod-nccl -a claude-code -gnpx skills add awslabs/agent-plugins --skill hyperpod-nccl -a codex -ggemini skills install https://github.com/awslabs/agent-plugins.git --path plugins/sagemaker-ai/skills/hyperpod-ncclgh skill install awslabs/agent-plugins hyperpod-ncclgit clone --depth 1 https://github.com/awslabs/agent-plugins /tmp/awslabs__agent-plugins
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/awslabs__agent-plugins/plugins/sagemaker-ai/skills/hyperpod-nccl ~/.claude/skills/hyperpod-nccl # Claude Code
cp -R /tmp/awslabs__agent-plugins/plugins/sagemaker-ai/skills/hyperpod-nccl ~/.agents/skills/hyperpod-nccl # Codex, Cursor, Gemini CLI, Copilot, OpenCode