gabriel / musehub public
filed by gabriel human · 38 days ago

Ownership/access follow-up: retire musehub-infra elsewhere, Namecheap/Cloudflare ownership, access review cadence

0 Anchors
— Blast radius
— Churn 30d
0 Proposals

Context

Split out from #150, which is now closed. #150 resolved most of what was originally blocked on AWS access: IAM Identity Center, individual identities, root MFA, and break-glass all landed. deploy/push.sh was also fixed to use individual SSO identity for staging deploys instead of the shared musehub-infra credential (verified live).

Remaining items

  • musehub-infra static credential still referenced elsewhere — push.sh no longer uses it, but these still do:
    • deploy/bootstrap-instance.sh — also has a stale PROD_INSTANCE ID from before the production rebuild
    • deploy/publish_muse_release.sh — references musehub-infra profile or default in its prerequisites comment; only targets staging, but not SSO-aware
    • deploy/backfill_loop.sh — targets staging instance directly, not SSO-aware
    • Once these are updated (or confirmed no longer needed), the musehub-infra IAM user's access key can actually be deactivated/deleted — right now it still exists as a live, usable shared credential even though the main deploy path no longer needs it.
  • Namecheap domain ownership — needs Gabriel to confirm musehub.ai is registered to a company account, not a personal login. Not checkable from this environment.
  • Cloudflare account ownership — same; needs Gabriel to confirm company-controlled.
  • Aaron's Cloudflare/DNS access — unverified. This is the one capability in #150's independent-operation test that's still an open question (Aaron can deploy, inspect logs, restore backups, and administer AWS independently now — DNS is the exception).
  • Quarterly access review process — doesn't exist as a recurring process yet, just point-in-time verification done during this sweep.
  • Ownership-per-area recording — AWS admin and deploys are now genuinely shared (Gabriel + Aaron). DNS/certs, DB admin, incident response, and billing still default to "Gabriel does everything" with no explicit second owner assigned to those specific areas.
◎ Activity
●
gabriel opened this issue 38 days ago
◎
No activity yet. Use the CLI to comment.