supported devices: add the VXE R1 on both transports

[SakethKanchi]

Adds the VXE R1 to the supported-devices table on both of its transports.

Depends on OpenMouse-Project/mouse-protocol#44, which makes the ATK driver
identify VXE mice by CID/MID and claim VXE's wired transport. Please merge that
first — until it lands (and openmouse's git dependency picks it up), the
supported claim here is only true against a local build.

Changes

• supported-mice.ts: one row, pids: [0x1085, 0xf58f] — the 2.4 GHz receiver and
  wired. Both are shared product ids, so the note points at the CID/MID
  identity (2,12, PAW3395) rather than implying the ids identify the model.
• supported-mice.test.ts:
  • VXE registered as a driver-reported brand, alongside the existing CRDRAKO
    entry — the ATK driver reports it via deviceBrand().
  • both product ids pinned in PID_UNIVERSE. The ATK discovery filter matches on
    vendor + usage page, so the protocol pins no ATK product ids of its own;
    0xf58f is pinned by ATK_COMPX_PRODUCT_IDS because it sits under the shared
    0x3554.

No driver or controller wiring was needed: AtkHidClient is already imported in
src/device/controller.ts and grouped in DM_CLASSES/DEDICATED.

Verification

npm run check passes (112 tests) with mouse-protocol#44 installed locally per
CONTRIBUTING's overlay recipe.

Verified in Chrome over WebHID on a real VXE R1 (firmware Mouse 3.13): the device
connects through the picker without crashing, names itself VXE R1, shows brand
VXE, 1,000 Hz, lift-off Low, Motion Sync on, ripple off, and angle snapping as
N/A (its angle register is unprogrammed). DPI tracked the mouse's own DPI-button
stage correctly — 1,200 DPI on stage 1, 1,600 on stage 2.

Battery also now shows a real charge state rather than "Unknown" — Charging on
the cable, Discharging on the receiver.

Before the protocol fix the same device showed ATK Wireless mouse -1k dongle at
320 DPI.

Note for a maintainer

ui/device-images.ts has no R1 asset and matches artwork by display-name regex, so
VXE R1 falls back to unknown-device.png. I deliberately did not map it to
another model's render.

[logm1lo]

Continued the R1 work in #163, depending on protocol PR OpenMouse-Project/mouse-protocol#53. #163 retains your original app commit in its history, narrows the claim to the physically tested wired R1 SE+, and adds Chromium WebHID read/write verification.