sagemaker-ai Plugin
Plugin Claude Code Development AI Agents & AI App DevelopmentCloud, Deployment & CI/CDThe explanation below is AI-generated. Please verify it against the sources.
sagemaker-ai is an agent plugin from the awslabs/agent-plugins repository that brings AWS AI/ML expertise directly into coding assistants for the surface area of Amazon SageMaker AI. According to its README, it provides 17 agent skills organized around two capability areas: model customization (guided fine-tuning of foundation models from use case definition through deployment) and HyperPod cluster operations (remote diagnostics and troubleshooting for training clusters). It also includes one MCP server, aws-mcp, for AWS documentation and SOP retrieval. The plugin can be installed in Claude Code or Cursor, or manually for other agents such as Kiro, and requires AWS credentials with appropriate IAM permissions to invoke SageMaker AI (and optionally Bedrock) operations.
Overview
Per the README, sagemaker-ai is not a standalone SaaS product but a plugin package combining agent skills, an MCP server, and reference documentation. It equips AI coding assistants to carry out Amazon SageMaker AI tasks such as fine-tuning foundation models and operating/debugging SageMaker HyperPod clusters.
What you can do with sagemaker-ai
- Build a step-by-step model customization plan using the planning skill
- Define use case goals, constraints, and success criteria with use-case-specification
- Validate and transform datasets for SageMaker-compatible training formats (dataset-evaluation, dataset-transformation)
- Select a fine-tuning technique (SFT, DPO, RLVR, etc.) and configure/run training jobs (finetuning-setup, finetuning)
- Design evaluations, select benchmarks, and use LLM-as-a-judge for model comparison (model-evaluation)
- Configure and launch endpoints on SageMaker AI or Amazon Bedrock (model-deployment)
- Run remote commands and transfer files on HyperPod nodes via SSM (hyperpod-ssm)
- Check and compare software component versions across HyperPod nodes (hyperpod-version-checker)
- Generate diagnostic issue reports for troubleshooting or support cases (hyperpod-issue-report)
- Diagnose cluster-wide, NCCL, per-node, performance, and Slurm scheduler issues on HyperPod clusters (hyperpod-cluster-debugger, hyperpod-nccl, hyperpod-node-debugger, hyperpod-performance-debugger, hyperpod-slurm-debugger)
Sources
Original description (English)
Build, train, and deploy AI models with deep AWS AI/ML expertise brought directly into your coding assistants, covering the surface area of Amazon SageMaker AI.
History of sagemaker-ai
- Plugin Added sagemaker-ai