[caumeira]

Resubmit of the pump RGB half of the earlier PR (#1, closed), addressing the review there.

Summary

• One Galahad2LightingDeviceProvider/Galahad2LightingFrameWriter now drives both the wired Trinity and the LCD variant, through a small IGalahad2PumpTransport (IsConnected + SendLighting) that Galahad2Hub and the LCD's JpegPanelHub each implement - only which object holds the HID handle differs. No second provider, no nullable fallbacks in the composite.
• Pump RGB rides the same 0x83 inner/outer/both command the wired Trinity already uses - confirmed against lian-li-linux, OpenRGB, and FanControl.LianLi, all of which describe this pump as 0x83-only.
• Dropped the earlier per-LED (0x14) path entirely: it had no citable source, so it's not shipped uncited. Can come back as a follow-up once there's something to cite (lian-li-linux doesn't have it; checked).
• LightingDevicesCatalog marks the Galahad II LCD row rgb-capable (previously screen-only).

Test plan

• dotnet build succeeds
• Verified on real Galahad II LCD hardware: canvas colours apply to the pump head through the two-ring path

I've read and agree to the CLA (CONTRIBUTING.md / CLA.md).

[caumeira]

Both points addressed:

One provider — folded the LCD into the existing Galahad2LightingDeviceProvider/Galahad2LightingFrameWriter. IGalahad2PumpTransport (IsConnected + SendLighting) is implemented by Galahad2Hub and by the LCD's JpegPanelHub; the provider picks whichever is connected (ActiveTransport). Deleted Galahad2LcdLightingDeviceProvider/Galahad2LcdLightingFrameWriter, and the _galahad2Lcd?.IsConnected ?? false / _galahad2Lcd! fallbacks in the composite are gone with them - there's nothing left to fall back from.

0x14 — took option 1: dropped it, shipped the LCD pump on the reference-backed 0x83 two-ring path instead (same command the wired Trinity already uses). Verified on my unit: canvas inner/outer colours apply to the pump head. Didn't end up needing the half-red/half-blue check since per-LED isn't in this PR anymore - if it comes back later I'll run that first and cite whatever it points to (SignalRGB's plugin source sounds like the right place to start).

Comments trimmed to one line each, no provenance narrative.

Also closed #1 and opened this fresh off main, per your note.

[caumeira]

Reopened as #5 with the structure fixed (the LCD's pump is one 12-LED ring, not the Trinity's two 1-LED rings this version assumed) and the 0x14 per-LED command restored with a real citation. GitHub wouldn't let me reopen this PR directly.