If you need to document how much AI actually touched your codebase, this gives you a structured way to do it. It's based on the AIAS v2 framework from education research, adapted for software projects. You get five levels from "no AI" to "experimental AI exploration," with specific indicators for each. Useful when you're writing transparency reports, creating disclosure badges, or just trying to honestly communicate to stakeholders what role Claude or Copilot played in your work. The framework won't certify quality or tell you if you're using AI "correctly," but it will help you be specific about where human judgment ended and AI generation began.
npx skills add https://github.com/mastepanoski/claude-skills --skill ai-assessment-scale