The N100 becomes the lived-in box (09 §3 decision 254): media apps installed and wired through the dashboard; films proven never off-site; the USB split NOT done (it is the whole-guest backup target)

- 09 §3 decision 254 (operator decisions 240/243 written down); target-selection: a "lived-in" tier, Tier 0 = demo-hp
  alone; nodes.md role + drive; 11-os-updates ring 0 wording; CLAUDE.md + docs.md: the diary.
- documentation/lived-in/n100.md (new diary); audits/n100-lived-in-2026-10-15/ (state, drive read, Part C capture sets,
  installs, wiring, stranger probes, demo-hp clean-up, the operator's SHEET.md).
- Register: R-1062 (the split, operator decides), R-1063 (decommission half-state on an agent refusal), R-1064 (path
  removal silent about cross-drive copies), R-1065 (a claimed Plex gets no libraries until a wiring run).
- demo-hp: /mnt/sizing-bench removed through the dashboard; the hub shows demo-hp OK (14:56 UTC).

Claude-Session: https://claude.ai/code/session_0159rPz1ZhFKsS53msqPYxtS

documentation/architecture/09-update-architecture.md

@@ -948,6 +948,26 @@ its length, and both fixes cost something the household would notice — operato
      **Extended the same day (operator, R-1059 = A):** the off-site and PBS-DR re-issues (they mint a household
      credential for the box) are session only too; „Off-site: write" keeps the windows and freeze.
 
+### 2026-10-15 (evening) — the N100 is the lived-in box (recorded by the n100-lived-in task; decisions 240 and 243)
+
+254. **The N100 (`demo-felhom`) is the lived-in box.** *Operator ruling 2026-10-14 (decision 240), its timing
+     decision 243 (after the Tester 1 all-drives test — done 2026-10-15), written down 2026-10-15 at the operator's
+     request.* Decision 240 said the rule „is recorded by the second task of 2026-10-14" — it never was; this is it.
+     1. **It changes only through the product's own paths:** the dashboard, the hub, the normal release train (the floor
+        under decision 250, the weekly golden), the OS update rings. No drill catalog, no hand edits inside the guest, no
+        `docker` command that changes state, no reinstall.
+     2. **No deliberate failure** — power cut, drive pull, drill, restore test — **without the operator's word for that
+        one act.** A failure that happens by itself is observed and recorded, and repaired by hand only where the
+        household could do the same.
+     3. **Reads, logs, metering and observation are free.** A problem found goes into the diary
+        (`documentation/lived-in/n100.md`) and the register.
+     4. **It stays a demo box:** no paying customer's data; it is not Tier 2. It is no longer „reach here first" — Tier 0
+        is demo-hp (and its guests and VMs) alone (`runbooks/target-selection.md`).
+     5. **The one hand change granted before the rule (operator, 2026-10-15 16:31): split the USB drive in two.** NOT
+        done: the drive is the box's whole-guest backup target (`felhom-backup`) and holds its nightly backups, so the
+        product refuses to decommission it and the N100 has no other drive to move that role to (R-1062,
+        `audits/n100-lived-in-2026-10-15/B/`). The grant is not used; a split needs a new word.
+
 ### 2026-10-15 (afternoon) — a restore puts the setup gate back as it was (operator brief, R-1057 raised to P2)
 
 253. **A restore puts the setup gate back exactly as it was when the backup was taken.** *Operator brief 2026-10-15
@@ -1008,6 +1028,8 @@ its length, and both fixes cost something the household would notice — operato
      (`15-media-library.md` §2; R-1014)
 243. **The N100 lived-in setup waits until the all-drives build has been tested on the Tester 1 rig.** Until then the
      N100 stays Tier 0 and is not fenced (it may host a second test rig). This replaces decision 240's timing.
+     **Fulfilled 2026-10-15** (the Tester 1 rig was tested; release 0.313.0): the rule is decision 254 — the N100 is no
+     longer Tier 0.
      *Operator ruling 2026-10-14.*
 
 ### 2026-10-14 (afternoon) — media apps on every drive (recorded before the work; the media-all-drives brief)
@@ -1020,7 +1042,8 @@ its length, and both fixes cost something the household would notice — operato
 239. **Each drive shows whether its apps have restarted since it was added.** If they have not, a manual restart
      button is there. *Operator ruling 2026-10-14.* (R-1023)
 240. **The N100 becomes a lived-in box.** Its rule is recorded by the second task of 2026-10-14
-     (`TASK-media-folder-fix-and-n100-2026-10-14.md`). *Operator ruling 2026-10-14.*
+     (`TASK-media-folder-fix-and-n100-2026-10-14.md`). *Operator ruling 2026-10-14.* **That task did not record it** (its N100 part was removed the
+     same evening, decision 243); the rule is decision 254 (2026-10-15).
 241. **Tester 1 may follow the drill catalog for the media test, until the operator says „tear it down".** It is the
      one exception to §6.5's fence; the teardown is R-1022. *Operator ruling 2026-10-14.*
 

documentation/architecture/11-os-updates.md

@@ -148,8 +148,9 @@ has not ruled on it as such.
 
 ### 5.1 Two rings
 
-- **Ring 0:** demo-felhom (N100) and demo-hp. They are disposable (`runbooks/target-selection.md`), and
-  they have different hardware. They take every update first.
+- **Ring 0:** demo-felhom (N100) and demo-hp, on different hardware. They take every update first. demo-hp is
+  disposable; the N100 is the lived-in box (`09` §3 decision 254) — it takes ring 0 through the product like any box,
+  which is the point: a lived-in box meets an update before a customer does (`runbooks/target-selection.md`).
 - **Ring 1:** every other box. It takes only what ring 0 approved.
 
 ### 5.2 Two lanes