Add z.ai/GLM provider via per-process Claude CLI env injection

Route glm:-prefixed models through the CLI/Anthropic path: ResolveModelId
strips the glm: prefix, and CreateCliStartInfo activates the z.ai
Anthropic-compatible endpoint (ANTHROPIC_BASE_URL/AUTH_TOKEN) only for that
process when IsZaiModel matches, leaving every other request's OAuth login
untouched. GLM model list is loaded through the existing model cache and
exposed via /api/models/glm with a "GLM (z.ai)" optgroup in the agent form.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Services/OpenAiService.cs

@@ -71,6 +71,7 @@ public class OpenAiService
     private List<ProviderModel> _geminiModels       = new List<ProviderModel>();
     private List<ProviderModel> _openaiModels       = new List<ProviderModel>();
     private List<ProviderModel> _minimaxModels      = new List<ProviderModel>();
+    private List<ProviderModel> _glmModels          = new List<ProviderModel>();
 
     // Fallback-Modell-Liste für MiniMax (falls /v1/models nichts liefert).
     // Quelle: https://platform.minimax.io/docs/api-reference/api-overview
@@ -126,6 +127,7 @@ public class OpenAiService
                     _geminiModels       = cache.ContainsKey("gemini")       ? cache["gemini"]       : new List<ProviderModel>();
                     _openaiModels       = cache.ContainsKey("openai")       ? cache["openai"]       : new List<ProviderModel>();
                     _minimaxModels      = cache.ContainsKey("minimax")      ? cache["minimax"]      : new List<ProviderModel>();
+                    _glmModels          = cache.ContainsKey("glm")          ? cache["glm"]          : new List<ProviderModel>();
                     // Falls der Cache älter als die MiniMax-Integration ist oder leer ist:
                     // statische Fallback-Liste verwenden, damit das Dropdown nicht leer bleibt.
                     if (_minimaxModels.Count == 0)
@@ -195,6 +197,13 @@ public class OpenAiService
             Console.WriteLine($"[OpenAiService] MinimaxApiKey nicht konfiguriert – nutze Fallback-Liste ({_minimaxModels.Count} Modelle).");
         }
 
+        // z.ai / GLM: Key aus Settings (ZaiApiKey). OpenAI-kompatibler /models-Endpoint von z.ai.
+        var zaiKey = _settings.Get("ZaiApiKey");
+        if (!string.IsNullOrEmpty(zaiKey))
+            _glmModels = await FetchModels("https://api.z.ai/api/paas/v4/models", "GLM (z.ai)", zaiKey);
+        else
+            Console.WriteLine("[OpenAiService] ZaiApiKey nicht konfiguriert – GLM-Modelle nicht geladen.");
+
         // Cache speichern
         try
         {
@@ -204,7 +213,8 @@ public class OpenAiService
                 ["pollinations"] = _pollinationsModels,
                 ["gemini"]       = _geminiModels,
                 ["openai"]       = _openaiModels,
-                ["minimax"]      = _minimaxModels
+                ["minimax"]      = _minimaxModels,
+                ["glm"]          = _glmModels
             };
             var json = JsonSerializer.Serialize(cache, new JsonSerializerOptions { WriteIndented = true });
             await File.WriteAllTextAsync(_cacheFile, json);
@@ -264,6 +274,7 @@ public class OpenAiService
     public List<ProviderModel> GetGeminiModels()       => _geminiModels;
     public List<ProviderModel> GetOpenAiModels()       => _openaiModels;
     public List<ProviderModel> GetMinimaxModels()      => _minimaxModels;
+    public List<ProviderModel> GetGlmModels()          => _glmModels;
 
     /// <summary>
     /// Gibt den Provider und das eigentliche Modell-ID aus einem kombinierten Model-String zurück.