Production readiness — Section 1: Ownership and Shared Access
aaronrene already has write collaborator access on musehub and muse repos. Everything else (AWS/Namecheap/Cloudflare ownership, recovery paths, break-glass procedure) needs Gabriel — requires AWS/Namecheap console access this environment does not have.
Key risk: deploys currently use a single shared AWS credential (musehub-infra), not individual identities. Aaron has no independent AWS identity yet — blocks the checklist's explicit "test that Aaron can operate independently" requirement.
Re-verified with full account access — most of this section is now resolved.
Resolved:
- IAM Identity Center + individual SSO identities for Gabriel and Aaron, each with MFA
- Aaron has AdministratorAccess on all three AWS accounts, matching his 50/50 ownership
- Root MFA on Management/Production; Nonproduction's root deliberately has no standing credential (centralized break-glass instead)
- Company-controlled AWS root mailboxes (Google Groups on the company domain)
- No more shared credential for deploys —
push.shnow uses individual SSO for both staging and production; verified live with a real staging deploy under Aaron's/Gabriel's own identity, zero reliance onmusehub-infra - Aaron upgraded from
writetoadmincollaborator onmusehubandmuse - Joiner/leaver procedure documented (identity model doc §11/§12)
- AWS break-glass procedure exists (centralized root access management)
Independent-operation test — 6 of 7 capabilities now genuinely true (sign-in, inspect prod, deploy staging, deploy/approve prod, inspect logs, restore backup). DNS/Cloudflare access is the one still-unverified capability.
Still open, split to #175: musehub-infra still referenced in a few other deploy scripts, Namecheap/Cloudflare ownership confirmation, Aaron's DNS access, quarterly access review cadence, per-area ownership recording (DNS/DB/incident-response/billing still default to Gabriel alone).
Closing — the core identity/access architecture this section was blocked on is done.
Scope update: deploy team expanded from 2 (Gabriel, Aaron) to 5 (+ interns Sol, Toby, Ben). Interns are staging-only by design. New runbook covering this: docs/production-readiness/runbook-deploy-onboarding.md. Interim: the current shared musehub-infra AWS credential should NOT be handed to interns before Section 3 ships individual identities.