olivercode-adoptionlisted

# OliverCode Adoption Drive the whole setup so the user does not have to read the documentation first. You read it; they answer questions. `$OC/docs/ADOPTING.md` is the reference contract — consult it when you need detail, but do not make the user read it. ## The rule that governs everything here **Optimize, do not amputate.** You are pointing existing rules at a shared home, not deleting them. A rule you cannot classify is a rule you keep and flag. Never delete a user's config because it looks redundant — redundancy is their call. Back up before the first mutation. Do not remove a backup until the user confirms the result. ## Run this before anything else ```bash bash scripts/qa/check-fork-usable.sh ``` If it fails, stop and report: the fork is broken, not their machine. Everything below assumes a sound checkout. ## The four decisions only a human can make Everything else is mechanical. Bring the user exactly these, with your recommendation and the evidence for it — do not decide them silently, and do not dump the whole inventory on them. **1. For each existing rules file: global, repo-specific, or personal?** This is the whole adoption. Most setups tangle all three in one file. - *machine-global* (applies everywhere) → the fork's `CLAUDE.md` - *repo-specific* (true of one product only) → that repo's `CLAUDE.md` - *personal/business* (voice, GTM, hiring) → their own user-skills area, NOT the framework - *cannot classify* → leave it exactly where it is and say so Pr