Use Veo by mode, not by name alone

Read video-provider-capabilities, video-cinematography, and exactly one provider skill: fal-ai for fal-ai/veo... routes or google-ai for Gemini API/Vertex. Then read the official docs for the exact API surface:

• fal Veo 3.1 Lite image-to-video: https://fal.ai/models/fal-ai/veo3.1/lite/image-to-video/api
• fal Veo 3.1 image-to-video: https://fal.ai/models/fal-ai/veo3.1/image-to-video/api
• Gemini API: https://ai.google.dev/gemini-api/docs/video
• Vertex first/last frames: https://cloud.google.com/vertex-ai/generative-ai/docs/video/generate-videos-from-first-and-last-frames
• Vertex prompt guide: https://cloud.google.com/vertex-ai/generative-ai/docs/video/video-gen-prompt-guide

Do not combine fal, Gemini, or Vertex model IDs and request shapes. Resolve the current model variant, availability, mode, limits, field nesting, and price on the surface actually configured for the production.