feat: implement error classification and handling for AI generation limits

- Added error classification for daily generation cap and concurrency limits in the PromptNode component, improving user feedback during AI image generation failures.
- Enhanced toast notifications to provide specific messages for daily limit and concurrent job errors.
- Introduced internal mutations in the credits module to check abuse limits and track usage, ensuring better resource management and user experience.
- Updated AI error handling logic to categorize and respond to different error types effectively.

lib/toast-messages.ts

@@ -45,6 +45,10 @@ export const msg = {
       title: "OpenRouter möglicherweise gestört",
       desc: "Mehrere Generierungen fehlgeschlagen.",
     },
+    concurrentLimitReached: {
+      title: "Generierung bereits aktiv",
+      desc: "Bitte warte, bis die laufende Generierung abgeschlossen ist.",
+    },
   },
 
   export: {