AI System Oversight Guide
Guides decisions about deploying, evaluating, and monitoring AI systems responsibly.
by cbrock84 · ★ 2,086 on GitHub
What it does
You describe an AI feature or model you're planning to use or already have running. Claude helps you define its intended use, how risky it is, what to test before launch, and how to keep watching it afterward. You get a clearer framework for deciding whether an AI system is safe and appropriate to use.
Who it’s for
A good fit for people who review contracts and legal documents, people who research markets, competitors or topics and people who run the day-to-day of a business.
Do you need to code?
No. You can use this without writing or editing any code.
How to install
Add it as a skill, or as part of its plugin. This skill also ships inside a plugin, so you can upload just this skill to Claude, or install the whole plugin it comes in. See the step-by-step guide.
Safety note
Gives legal/financial/medical info - not professional advice.
As with any add-on, only install skills from sources you trust, and read what it does first.
License
MIT