I've been finding general direction prompting for dialogue is getting better results than giving specific lines. Text is less likely to show up on the screen and the actor flubs their lines less.

image:
ultra-detailed skin texture. High-end hyperrealistic photo with sharp focus, 8k quality, masterpiece, (best quality:1.3), (highly detailed skin:1.2), (volumetric lighting:1.1), (Scandinavian:1.5), (very slender). A stunningly beautiful young very slender very young female catgirl, 18 years old, long curly pink hair that goes down to her waist, fuzzy ears and tail like a cat's, huge beautiful expressive green eyes with long lashes, delicate facial features, girly makeup, pale and freckled skin, massive breasts, cleavage with visible nipples, white collar with gold bell, gold princess tiara, laced up pink corset with a white skirt, white stockings held up by garters, pink high heels, long white gloves that go past her elbow, slim toned narrow waists, toned athletic abdomen. Shot in a grassy field at the edge of a medieval town at night. She is tied to a big wooden X that is on top of a massive two meter tall pile of firewood ready to light. Her arms and legs are stretched out straight, following the cross. She is tied to the crossed pieces of wood by her ankles and wrists with hempen rope. High angle view. Cinematic lighting.


video:
music=none. mode=spicy.
a torch is thrown into the pile of wood, landing lit end first at the bottom of the pile. The catgirl hyperventilates and panics as a flame under her grows bigger and bigger. her arms and legs stay tied to the cross. She is burned alive, screaming in pain about how much it hurts begging someone to save her. Her flesh blisters and skin melts off. Her clothes and hair burn.