# Part 5 — claims walked (2026-11-02/03), on Tester 1, front doors only; mail read in the mailbox

| Claim | Result | File |
|---|---|---|
| fail.app-crash, fault.customer-email, fault.operator-email | walked | `crash-walk*.log`, `mail-*.txt` |
| backup.fill-warning, fail.drive-filling | walked | `fill-walk.log`, `mail-fill-household.txt` |
| backup.restore-proof | walked (journals of three boxes) | `../part4/restore-test-*.txt` |
| backup.offsite | walked (snapshots per day) | `offsite-*.txt` |
| fail.internet-down | walked; slow Backups page, R-1362 | `outage-wan.log` |
| use.lan-fallback | partial; R-1364 | `lan-resolver.txt` |
| fail.hub-down | partial; events lost, R-1361 | `outage-hub.log` |

Tries that taught something, kept: `outage-wan-try1-…` (the hub walk's own safety job lifted the block mid-way), `outage-wan-try2-…` (removing the default route: the agent's self-heal put it back in about 100 s; a good thing, and a wrong way to stage a dead line).

Left on Tester 1: `notifications.email_cleared: true` in its settings (the address was empty when found, typed for the walk, cleared through the page).