sync: auto-sync from HOWARD-HOME at 2026-07-05 21:24:32

Author: Howard Enos
Machine: HOWARD-HOME
Timestamp: 2026-07-05 21:24:32

.claude/SKILL_ROUTING.md

@@ -23,11 +23,7 @@
 | Run a command / investigate / script on an RMM agent | `/rmm` (find the host first with `rmm-search`) |
 | M365 investigation/remediation, breach check, mailbox/inbox-rule/forwarding audit, tenant sweep | `remediation-tool` |
 | Onboard a new M365 tenant to the remediation app suite | `onboard365` |
-| Is client X backing up? backup status/failures/stale plans; MSP360 backup user/company/license provisioning | `msp360` (authoritative backup layer — check here, not bucket contents) |
-| Backblaze / B2 storage, buckets, backup storage cost (destination side) | `b2` |
-| Who backs up to Seafile / ownCloud / Datto Workplace (per-backend inventory + endpoint detection) | `seafile` / `owncloud` / `datto-workplace` |
-| Provision/admin Seafile (SeaCloud) — create/deactivate/delete account, set quota, create/transfer/delete library, share/unshare | `seafile` (writes preview by default; `--confirm` to apply) |
-| Provision/admin ownCloud (cloud.acghosting) — create/delete/enable/disable user, reset password, set quota, groups, apps/config, maintenance mode, files scan/transfer, trashbin/versions cleanup | `owncloud` (writes gated `--confirm`; `--all-users` purge also needs `--force-all-users`) |
+| Anything about a customer's **BACKUP** — is X backing up, who backs up where, backup status/failures, or set up / provision a backup | **Backups → wiki-first** (see the "Backups" block right below this table) |
 | Bitdefender / GravityZone AV — endpoints, sweeps, policies, quarantine, EDR | `bitdefender` |
 | Datto EDR / Datto AV — detections, isolate, scan, agent deploy | `datto-edr` |
 | VoIP — PacketDial / OITVOIP / NetSapiens domains, users, DIDs, queues, CDRs | `packetdial` |
@@ -50,6 +46,31 @@
 
 > Not listed? If no skill fits, do it directly — and say so. New recurring need → `skill-creator`.
 
+### Backups (wiki-first routing)
+
+A customer can back up through **any** of several backends, so a backup request is a
+two-step route: **first find out which backend THIS customer uses, then use that skill.**
+Never assume B2/MSP360 — several clients are on Seafile, ownCloud, or Datto Workplace.
+
+1. **Identify the backend for this customer (don't guess).** Read the client wiki
+   (`wiki/clients/<slug>.md`) and the backup map (`projects/gps-rmm-audit/backup-map.md`) —
+   both record each customer's real destination. If it isn't recorded, confirm with
+   `msp360 monitoring --company <name>` plus the per-backend inventory skills, then write the
+   answer back to the client wiki.
+2. **Route to that backend's skill:**
+
+   | Need | Skill |
+   |---|---|
+   | Is X backing up? status / last run / failures / stale plans (MSP360 + B2 Cloud plans) | `msp360` — **authoritative**, check here not bucket contents |
+   | Who backs up to Seafile / ownCloud / Datto Workplace (inventory + endpoint detection) | `seafile` / `owncloud` / `datto-workplace` |
+   | Storage / bucket / cost on the destination side | `b2` (Backblaze) |
+   | **Set up / provision a backup** — create account/user, quota, library, license, destination | the **target backend's** skill: `msp360` (MBS user/company/license), `seafile`, `owncloud` (all writes gated `--confirm`); `b2` for a new bucket/key |
+
+- **Which machine gets backed up is a per-client decision** — report mismatches (billed vs
+  running), don't provision jobs off a billing count. See [[feedback_backup_targets_never_guessed]].
+- Full per-customer destination + health map: `projects/gps-rmm-audit/backup-map.md`;
+  how backup-status works: memory `[[reference_msp360_backup_monitoring]]`.
+
 ---
 
 ## Check-skills (work-type → the gate run BEFORE "done")