Cookbook UI: backend-aware env vars, always-show MoE/EP/Reasoning toggles, GPU default, Firefox-mobile expand

Frontend half of the backend-detection + per-OS install command work,
plus a pile of mobile/UX fixes:

Backend awareness:
- _gpuEnvPrefix() picks CUDA_VISIBLE_DEVICES / HIP_VISIBLE_DEVICES /
  nothing based on detected hwfit backend + scanned-host match (so a
  stale ajax scan does not leak CUDA env vars into a kierkegaard
  Vulkan launch). Replaces 6 hardcoded CUDA_VISIBLE_DEVICES sites.
- GGML_CUDA_ENABLE_UNIFIED_MEMORY only emitted when backend is
  actually CUDA (was leaking onto Vulkan/ROCm via saved presets).

Per-target install command:
- Dep rows render a single mono command box + Copy button when the
  server resolved pkg.install_cmd_for_target. Reused in the build-deps
  install failure toast so the toast and the row show the same line.
- Diagnosis patterns split cmake/g++/git out of the generic
  llama-cpp-python catch-all so a missing-cmake failure surfaces a
  cmake-specific message + per-distro Copy buttons.

Form toggles always visible:
- Reasoning Parser, Expert Parallel, MoE Env Vars no longer gated on
  model-family detection. Detection still hints (parser tag shown when
  matched); toggle works with sensible defaults otherwise. MiniMax M-
  series added to MoE family detector so the auto-fill is right.

Mobile + GPU default:
- Launch tab cached-list flex collapsed to 0px on mobile because the
  desktop `flex: 1 1 0` had no parent height to grow into. Override
  to `flex: 0 0 auto` in the cookbook mobile @media block.
- doclib-card expand on mobile (Firefox no :has() support) pins
  explicit px heights so the launch form actually appears.
- llama_mode defaults to gpu when hwfit detected cuda/rocm/vulkan/
  metal on the current target, instead of always cpu (which was
  forcing -ngl 0 on first-open and burning 35GB models on CPU).

static/js/cookbook-diagnosis.js

@@ -461,6 +461,40 @@ export const ERROR_PATTERNS = [
       { label: 'Copy install command', action: () => _copyText('curl -fsSL https://ollama.com/install.sh | sh') },
     ],
   },
+  // System build deps must be checked BEFORE the llama-server catch-all:
+  // a `cmake: command not found` failure ALSO produces `llama-server:
+  // command not found` later in the script (the build aborts then the
+  // run line fails) — pattern order is first-match-wins, so without
+  // these specific entries the user gets the misleading "install
+  // llama-cpp-python[server]" suggestion when the actual blocker is a
+  // missing OS-package toolchain that pip can't ship.
+  {
+    pattern: /cmake: command not found|cmake.*not found.*Could not/i,
+    message: 'cmake is required to compile llama.cpp from source, but it is not installed on this server.',
+    suggestion: 'Suggested action: install cmake via the OS package manager — apt: cmake build-essential / pacman: cmake base-devel / dnf: cmake gcc-c++ make / brew: cmake. Cookbook can do this automatically on the next launch if your user has passwordless sudo for apt/pacman/dnf.',
+    fixes: [
+      { label: 'Open Dependencies', action: () => _openCookbookDependencies('llama_cpp') },
+      { label: 'Copy apt install', action: () => _copyText('sudo apt install -y cmake build-essential git') },
+      { label: 'Copy pacman install', action: () => _copyText('sudo pacman -Sy --needed cmake base-devel git') },
+      { label: 'Copy dnf install', action: () => _copyText('sudo dnf install -y cmake gcc gcc-c++ make git') },
+    ],
+  },
+  {
+    pattern: /^(make|g\+\+|gcc): command not found|Could not find C\+\+ compiler/i,
+    message: 'A C/C++ compiler (build-essential / base-devel) is required to compile llama.cpp.',
+    fixes: [
+      { label: 'Open Dependencies', action: () => _openCookbookDependencies('llama_cpp') },
+      { label: 'Copy apt install', action: () => _copyText('sudo apt install -y build-essential') },
+    ],
+  },
+  {
+    pattern: /^git: command not found/i,
+    message: 'git is required to clone the llama.cpp source tree.',
+    fixes: [
+      { label: 'Open Dependencies', action: () => _openCookbookDependencies('llama_cpp') },
+      { label: 'Copy apt install', action: () => _copyText('sudo apt install -y git') },
+    ],
+  },
   {
     pattern: /llama-server.*command not found|llama\.cpp.*not found|No module named.*llama_cpp|No module named 'starlette_context'/i,
     message: 'llama-cpp-python server is not installed. Run: pip install "llama-cpp-python[server]"',
@@ -688,11 +722,15 @@ export function _showDiagnosis(panel, diagnosis, sourceText) {
   copyBtn.addEventListener('click', async (e) => {
     e.stopPropagation();
     const bundle = _diagnosisCopyBundle(task, diagnosis, sourceText, suggestionText);
-    try {
-      await navigator.clipboard.writeText(bundle);
+    // Use the shared helper which falls back to execCommand('copy') on
+    // non-HTTPS origins (Tailscale IPs, LAN IPs, etc.) — navigator.clipboard
+    // is silently a no-op on those, which is why the button appeared dead
+    // for users on http://100.113.161.2:7011 over Tailscale/mobile.
+    const ok = await _copyText(bundle);
+    if (ok) {
       copyBtn.classList.add('copied');
       setTimeout(() => { if (copyBtn.isConnected) copyBtn.classList.remove('copied'); }, 1200);
-    } catch (_) {}
+    }
   });
 
   const dismissBtn = document.createElement('button');
@@ -757,7 +795,7 @@ export function _showDiagnosis(panel, diagnosis, sourceText) {
       });
       row.appendChild(btn);
     }
-    body.appendChild(row);
+    diag.appendChild(row);
   }
 }