AWSBestPracticesSkill
Unofficial source-linked AWS best practices for every AWS service, organized by Well-Architected pillar for Claude Code, OpenAI Codex, and AI coding agents.
Install
claude plugin install aws-best-practices@aws-best-practices-skillInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →AWS Best Practices Skill
Ask your AI coding agent "how should I secure my S3 bucket" and get sourced, locally cataloged AWS best practices, skipping both stale training data and costly live research.
A skill for Claude Code and OpenAI Codex that collects the best practices of every AWS service, and nothing else. Find recommendations by use case: security, reliability, performance, cost, operations, sustainability, each one linked to its official AWS source.
Project page: https://ferdinandobons.github.io/AWSBestPracticesSkill/
Latest release: v0.1.5
Unofficial project: this is an independent community-maintained skill. It is not an official AWS skill, AWS product, or AWS-maintained resource, and it is not affiliated with or endorsed by Amazon Web Services.
Scope: this skill contains only best practices. No service descriptions, no pricing, no tutorials, no code walkthroughs. Just what AWS recommends you do, organized so you can act on it, each item linked to its official AWS source.
Why this instead of just asking the model directly
There are three ways an AI agent can answer "what are the best practices for this AWS service":
- From memory: free and instant, but the model's training data goes stale the moment AWS ships a new feature, renames a service, or updates its Well-Architected guidance. It will still answer confidently even when it's wrong or out of date.
- Live research: the agent runs web searches or queries an AWS documentation MCP server on the spot, every time. This gets it right, but it's expensive: several search → fetch → read round-trips, burning tens of thousands of tokens, repeated for every question, even the same one asked twice.
- This skill: the research is already done, once, per service, and
stored as a small, source-linked file. The agent opens exactly
services/<category>/<service>.md, reads a few hundred lines, and answers, no live search needed and no repeated token cost, while every practice is still traceable to an officialdocs.aws.amazon.com/aws.amazon.com/wa.aws.amazon.compage. And because content ages, a documented refresh loop (below) keeps it from silently drifting back into "stale training data" territory.
What it contains / does NOT contain
| ✅ Contains | ❌ Does not contain |
|---|---|
| Best practices per service, by Well-Architected pillar | Service overviews / "what is X" |
[when-it-applies] context tags per practice |
Pricing, cost estimates, calculators |
| Cross-service general best practices | Tutorials / getting-started / how-to |
| Official AWS source link on every item | Extended code samples |
Quick start
Claude Code (plugin, recommended):
claude plugin marketplace add ferdinandobons/AWSBestPracticesSkill
claude plugin install aws-best-practices@aws-best-practices-skill
Update with claude plugin update aws-best-practices, remove with claude plugin uninstall aws-best-practices. (Same steps work as /plugin marketplace add ... and /plugin install ... from inside an active Claude Code session.)
Similar plugs
Multi-harness agentic plugin marketplace for Claude Code, Codex, Cursor, OpenCode, GitHub Copilot, Google Antigravity, and Pi
A Claude Code plugin that shows what's happening - context usage, active tools, running agents, and todo progress