infinite talk prompt question

[doktourtv]

Running comfyui portable on i9-14 64 Gig Ram RTX 5080 gpu

I am using infiinite talk (single). The workflow I am testing is the one provided in the instructional video. It is using WanVideo TextEncode Cached for the text input, however it is ignoring all instructions in the positive prompt. For example “add people in the background” or “The man is sitting” or “he is standing in front of a building” , all ignored. the render syncs the image to audio tract perfectly but I am unable to modify the image in any way using the prompt. Can someone suggest what I might be doing wrong?