feat: pos system backup copies to USB after verification — sha256-proven 100%

Once the archive verifies, USB detection runs (so a stick plugged in while
the backup ran is found): mounted removable storage is auto-detected via
lsblk -J + a recursive jq filter (rm, mounted, type part|disk — JSON makes
spacey mountpoints safe), or BACKUP_USB_ROOT pins a fixed stick and skips
detection. None mounted → one re-scan prompt ('s' skips, EOF from cron
skips silently, rc stays 0); one stick → y/N confirm; several → numbered
pick (0 = skip). The copy lands in <usb>/backups/ (mkdir -p, chmod 600
best-effort — a vfat chmod failure warns, never fails the copy) and the
transfer is proven 100% by sha256 source-vs-copy before any success is
announced; a mismatch warns with both hashes, notifies 'USB copy FAILED',
and exits 1. The ERR trap is re-armed before the USB phase so a copy
failure no longer notifies 'Backup FAILED'. Docs: usage() Environment,
POS.md backup row, howto/system.md (USB section + env table + mismatch
troubleshooting), DEV.md system.env list. Stub suite
(/tmp/opencode/backup-test, HOME-isolated, sudo/gpg/lsblk/sender stubs,
corrupting-cp + vfat-chmod overrides, per-test lsblk JSON fixtures):
40/40 green.

DOC/POS.md

@@ -212,7 +212,7 @@ The standalone `vbox` command still works and forwards to `pos docker vbox` (see
 | Command | File | Purpose | Configuration |
 |---------|------|---------|---------------|
 | `sudo pos system firewall` | `bin/pos-system-firewall` | Interactive UFW ("UFW POWER") menu: add/delete rules, status, enable/disable/reset, default policies | Must run as root. Every command is previewed and confirmed before execution; supports `--dry-run`; keeps a history of executed commands. Executed mutating changes are announced via `lib/notify.sh` |
-| `pos system backup <folder-path>` | `bin/pos-system-backup` | Create a gpg-encrypted (AES-256) `tar.gz` snapshot of a folder and verify it | Prompts twice for a password (never stored). Uses `sudo tar`; needs `gnupg` (in `preinstall.sh` PACKAGES). Artifact `<name>_<date>.tar.gz.gpg` in the current directory, `chmod 600`. Success/failure are announced via `lib/notify.sh` |
+| `pos system backup <folder-path>` | `bin/pos-system-backup` | Create a gpg-encrypted (AES-256) `tar.gz` snapshot of a folder and verify it | Prompts twice for a password (never stored). Uses `sudo tar`; needs `gnupg` (in `preinstall.sh` PACKAGES). Artifact `<name>_<date>.tar.gz.gpg` in the current directory, `chmod 600`. Once the archive verifies, connected USB storage is offered (auto-detected removable mounts via `lsblk`, or pinned with `BACKUP_USB_ROOT`): the copy lands in `<usb>/backups/` and is proven 100% by sha256 before it is announced. Success/failure are announced via `lib/notify.sh` |
 | `pos system backup --service` | `bin/pos-system-backup` | Lists folders under `/srv` and `~/srv`, lets you pick one, then runs the same backup | Roots via `BACKUP_SERVICE_ROOTS` (space-separated, default `/srv $HOME/srv`) or `~/.config/linux_post_install/system.env` |
 | `pos system health [--send] [--markdown]` | `bin/pos-system-health` | Host health dashboard: disk per mount, RAM/swap, failed systemd units, backup age, fail2ban, docker containers. Exits 1 if any check FAILs | `--send`/`--markdown` are notification-only: they send the summary via `lib/notify.sh` to every platform in `NOTIFY_PLATFORM` and do NOT print the dashboard (so wrappers like the Telegram listener don't echo it back — pair with the listener's `@quiet` marker). `HEALTH_BACKUP_MAX_AGE_DAYS` (default 2) and `BACKUP_SERVICE_ROOTS` come from `~/.config/linux_post_install/system.env`; `--help` shows the effective values. Platform list from `~/.config/linux_post_install/notify.env` |
 | `pos system schedule <cmd>` | `bin/pos-system-schedule` | Scheduled jobs — run a command on a timer, notify (or stay silent): `run [name\|all]`, `list`, `config`, `enable [name\|all]`, `disable [name\|all]`, `status`, `migrate`. Each job is a file in `~/.config/linux_post_install/schedule.d/<name>.env` with `INTERVAL` (`5m…59m`, `1h…23h`, `hourly`, `daily`, `weekly`, `OnCalendar=…`), `NOTIFY` policy, optional `MSG`, `RULE` (threshold only), and `COMMAND` = the literal rest of the line (pipes/quotes/`sudo` fine). Policies: `always` (full output every run), `onchange` (send when output differs from the last run; first run always sends), `onerror` (non-zero exit or empty output), `threshold` (first numeric output vs `RULE`, alert on false→true + one recovery — the old event-trigger behavior), `never` (side-effect jobs, no notify) | One systemd **user** timer pair per job (`pos-schedule-<name>.timer` + oneshot `.service`, `Persistent=true`), reconciled on `enable`/`disable`; the legacy single `pos-event-trigger` timer is auto-removed. `migrate` converts a pre-existing `event.env` rule set into `schedule.d/rule-N.env` threshold jobs. `config` is an interactive editor (add/edit/remove/enable/disable, validates interval + threshold); alerts via `lib/notify.sh`; `--dry-run` previews runs/writes/sends; jobs are arbitrary shell commands (chmod 600, same trust model as the Telegram map); starter jobs in `config/schedule.d/` auto-installed no-clobber by postinstall |