This walks you through Amazon S3 Vectors, AWS's cost-effective vector storage service for embeddings you don't need to query constantly. It covers the full workflow from creating buckets and indexes to storing and querying vectors, with clear decision points on when to use OpenSearch instead (high QPS, hybrid search). The skill is thorough about immutable configuration gotchas like dimension and distance metric choices, and it handles the Bedrock integration for generating embeddings. Good for RAG systems where you need cheap long-term storage with occasional queries. The troubleshooting table actually covers the common IAM permission trip-ups with the s3vectors namespace.
npx skills add https://github.com/aws/agent-toolkit-for-aws --skill storing-and-querying-vectors