aws-samples/sample-well-architected-skills-and-steering — reverse-engineered prompt

Build me a reusable starter repo that teaches AI coding assistants how to apply the AWS Well Architected Framework while they’re writing code. I want one shared set of guidance and playbooks that can be installed into tools like Cursor, Claude Code, Codex, Copilot, Cline, Gemini CLI, Windsurf, and similar agents, so teams get the same advice everywhere.

Make it work locally with no AWS credentials or API calls. Include always on guidance for the framework and a deeper review workflow, plus practical skills for security, reliability, cost, performance, sustainability, operational excellence, migration readiness, architecture decisions, and a builder that can help create diagrams or roadmaps. Add install scripts for Mac, Linux, and Windows that can auto detect supported tools, install into a specific project or globally, and optionally keep things linked for easy updates.

Please also include a couple of intentionally flawed serverless sample apps for demos and testing, and a simple evaluation runner to score how well an agent follows the guidance. Look up current docs online if you need to.