Vouch request: sdarive

[sdarive]

What do you want to work on?

I'm proposing #3997 as an example for the /examples folder. This is based on my experience in implementing AI inside enterprise environments.
This proposal demonstrates five controls that are determined by typical enterprise AI policy and enforced at run time:

• Control 1: least privilege
• Control 2: credentials
• Control 3, need to know
• Control 4, human approval, happens outside the file/prompt
• Control 5, approved vendors

This is already written and runs cleanly on 0.1.2, so the PR is small and self-contained (docs only, no code changes)

Why this change?

This is based on my experience of seeing way too many AI pilots fail because they can't satisfy risk and governance criteria with trust and auditability without adding to their burden. I want to show with an example to educate both AI developers and risk managers. I have also built a live demo using OpenShell at https://shikologic.com/demo to help

Checklist

• I wrote this request myself (not AI-generated)
• I have read CONTRIBUTING.md