refactor(validation): deduplicate image optimization, fix quality-reduction-before-chunking, unify chunk_count key

- Extract _optimize_images() helper to eliminate duplicate
  optimization code (was duplicated between initial pass and
  quality-reduction fallback)
- Move quality-reduction estimate to only apply when NOT chunking
  (each chunk fits the image limit by definition)
- Fix _merge_chunk_results to return 'chunk_count' instead of 'chunks'
  for consistency with plugin.py
- Simplify plugin.py: analysis.get('chunk_count', 1) instead of
  fallback chain
- Document that Kilo API gateway is incompatible with AsyncOpenAI
  image format (probe returns 0)

backend/src/api/routes/llm.py

@@ -481,6 +481,9 @@ async def probe_max_images(
     from openai import AsyncOpenAI
 
     # Minimal 1x1 white JPEG in base64 (~840 chars, PIL-generated)
+    # NOTE: Uses raw AsyncOpenAI client — works for OpenAI/OpenRouter providers.
+    # Kilo API gateway does NOT support the OpenAI image content format;
+    # probes against Kilo providers will return max_images=0 (not a bug).
     PROBE_IMAGE_B64 = "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAFA3PEY8MlBGQUZaVVBfeMiCeG5uePWvuZHI////////////////////////////////////////////////////2wBDAVVaWnhpeOuCguv/////////////////////////////////////////////////////////////////////////wAARCAABAAEDASIAAhEBAxEB/8QAHwAAAQUBAQEBAQEAAAAAAAAAAAECAwQFBgcICQoL/8QAtRAAAgEDAwIEAwUFBAQAAAF9AQIDAAQRBRIhMUEGE1FhByJxFDKBkaEII0KxwRVS0fAkM2JyggkKFhcYGRolJicoKSo0NTY3ODk6Q0RFRkdISUpTVFVWV1hZWmNkZWZnaGlqc3R1dnd4eXqDhIWGh4iJipKTlJWWl5iZmqKjpKWmp6ipqrKztLW2t7i5usLDxMXGx8jJytLT1NXW19jZ2uHi4+Tl5ufo6erx8vP09fb3+Pn6/8QAHwEAAwEBAQEBAQEBAQAAAAAAAAECAwQFBgcICQoL/8QAtREAAgECBAQDBAcFBAQAAQJ3AAECAxEEBSExBhJBUQdhcRMiMoEIFEKRobHBCSMzUvAVYnLRChYkNOEl8RcYGRomJygpKjU2Nzg5OkNERUZHSElKU1RVVldYWVpjZGVmZ2hpanN0dXZ3eHl6goOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKmqsrO0tba3uLm6wsPExcbHyMnK0tPU1dbX2Nna4uPk5ebn6Onq8vP09fb3+Pn6/9oADAMBAAIRAxEAPwC7RRRQB//Z"
 
     service = LLMProviderService(db)