## What happened

Locking the machine (manually via `omarchy system lock`, or via idle timeout)
never powers the displays off. The lock screen stays fully lit indefinitely.

## What was expected

`omarchy-shell`'s lock plugin arms a blank timer that runs
`omarchy-brightness-keyboard off; omarchy-brightness-display off` ~5 s after
locking (`shell/plugins/lock/Service.qml`, `blankProcess`), so the panels
should go dark shortly after lock.

## Root cause

`bin/omarchy-brightness-display` dispatches DPMS like this (current master,
lines ~63 and ~70):

```bash
hyprctl dispatch 'hl.dsp.dpms({ action = "disable" })'   # off branch
hyprctl dispatch 'hl.dsp.dpms({ action = "enable" })'    # on branch
```

On this system the `{ action = "disable" }` form returns `ok` but does
nothing — every output still reports `dpmsStatus: true` afterwards. The
string form works:

```
$ hyprctl dispatch 'hl.dsp.dpms({ action = "disable" })'; sleep 1
$ hyprctl monitors -j | jq '.[].dpmsStatus'
true
true

$ hyprctl dispatch 'hl.dsp.dpms("off")'; sleep 1
$ hyprctl monitors -j | jq '.[].dpmsStatus'
false
false
```

(`hl.dsp.dpms({ action = "enable" })` does relight the panels, which is
consistent with the dispatcher treating any unrecognized argument as "on" —
that would make the broken "off" branch invisible everywhere except a lock
that never blanks.)

## Suggested fix

Use the string form in `omarchy-brightness-display`:

```bash
hyprctl dispatch 'hl.dsp.dpms("off")'
hyprctl dispatch 'hl.dsp.dpms("on")'
```

## Steps to reproduce

1. `omarchy system lock` (or wait for idle lock)
2. Don't touch anything for a minute
3. Displays stay lit; `hyprctl monitors -j | jq '.[].dpmsStatus'` (from a TTY
   or over SSH) shows `true` on every output

## System

- Omarchy 4.0.0-1 (also reproduced against current master's
  `bin/omarchy-brightness-display`)
- Hyprland 0.56.2-1 (v0.56.2, commit efb5099)
- CPU: AMD Ryzen 9 7950X3D; GPU: NVIDIA RTX 4060 Ti (AD106) + AMD Raphael iGPU
- Monitors: 2× Dell SE3223Q (HDMI-A-1 + DP-3), scale 1.6
- Full `omarchy debug` log available on request

Possibly related but distinct: #7399 reports the lock screen *re-blanking*
after DPMS wake, which implies blanking does work on that setup — so this may
be version- or path-dependent rather than universal.