feat(ui): localize commit AI and git error messages to English

Translate commit AI prompts, model labels, and git/keychain errors to
English so the app and generated messages are consistent. Also add a
discard confirmation dialog and update the UI text/styles to match the
new flow.

- src-tauri/crates/commit_ai/src/cloud.rs
  - Translate HTTP and API error messages to English.
  - Keep request timeout and token sizing behavior unchanged.
- src-tauri/crates/commit_ai/src/lib.rs
  - Translate model labels, prompt text, and validation errors.
  - Keep diff truncation and message sanitization logic intact.
- src-tauri/src/git.rs
  - Translate git, credential, merge, and history errors.
  - Update AI provider validation messages to English.
- src/lib/components/*
  - Update AI settings, commit panel, credential, and loading UI text.
  - Add discard confirmation dialog for destructive actions.
- src/App.svelte, src/app.css
  - Adjust app layout and styling for the new dialog and text changes.

src/lib/components/AiSettingsDialog.svelte

@@ -153,9 +153,9 @@
         <div class="cred-token-hint">
           <AlertCircle size={13} aria-hidden="true" />
           <span>
-            Beim Wechsel wird das Modell{selectedLocalModel ? ` (${formatSize(selectedLocalModel.approx_size_mb)})` : ""}
-            im Hintergrund heruntergeladen — je nach Internetverbindung kann das mehrere Minuten dauern.
-            Danach bleibt es lokal zwischengespeichert und lädt beim nächsten Start sofort.
+            Switching downloads the model{selectedLocalModel ? ` (${formatSize(selectedLocalModel.approx_size_mb)})` : ""}
+            in the background — depending on your internet connection this can take several minutes.
+            After that it stays cached locally and loads instantly on the next start.
           </span>
         </div>
       {:else if provider === "openai"}
@@ -230,7 +230,7 @@
         </div>
         <div class="cred-token-hint">
           <Globe size={13} aria-hidden="true" />
-          <span>Für lokale OpenAI-kompatible Server wie Ollama oder LM Studio. Die Basis-URL sollte auf /v1 enden.</span>
+          <span>For local OpenAI-compatible servers like Ollama or LM Studio. The base URL should end in /v1.</span>
         </div>
       {/if}