amazon-aurora-postgresql

Amazon Aurora PostgreSQL is a modular AWS toolkit that helps an agent create, modify, and advise on Aurora PostgreSQL clusters specifically. It is organized as a registry of sub-skills, each handling one domain — cluster creation, express (single-API-call) configuration, Aurora serverless advisory, I/O-Optimized storage evaluation, commitment pricing, and PostgreSQL upgrade planning. A router matches user intent to the right sub-skill and loads only the reference files needed, so the agent works from the toolkit's own instructions rather than general knowledge. It explicitly defers Aurora MySQL work to a separate skill.

The skill emphasizes safety and read-first behavior. Its operating procedure routes, loads the matched instructions, analyzes or advises, and only then — if a mutation is requested — classifies the change against safety guardrail tiers, confirms with the user, applies resource tags, and executes (preferring the audited AWS MCP server, falling back to CLI). Commitment-pricing workflows are estimate-only: purchase APIs are blocked entirely, and the guidance repeatedly warns against forcing analyses on empty clusters, over-committing serverless capacity, or recommending long terms for workloads that may be retired. Express configuration provisions serverless with IAM-only authentication (no master password), connecting via short-lived IAM auth tokens. Bundled Python scripts (ACU calculator, I/O-optimized analyzer, commitment-pricing analyzer) perform the non-trivial cost and sizing math consistently.

It targets AWS engineers, database administrators, and cloud cost owners who need guided, guardrailed help with Aurora PostgreSQL provisioning, capacity planning, storage-type decisions, reserved-instance and savings-plan estimation, and version-upgrade planning.