A lone knight faces a dragon army before a ruined castle

2x2 grid, do this for 4 ai created characters, 16:9 input: protagonist = [any subject] opposing_force = [any threat / group / system] environment = [any location] action_beat = [any action or mood] run cinematic_showdown_extraction: protagonist_dna := infer age category, personality, silhouette, outfit, emotional tone, style language opposing_force_dna := infer number, formation, costume, scale, threat level, behavior environment_dna := infer location, destruction level, atmosphere, lighting, debris, terrain contrast_system := infer contrast between protagonist and opposing force action_state := infer whether scene is before impact, during action, or after impact camera_language := infer wide shot, close-up, dolly angle, handheld energy, or anime cinematic framing threat_geometry := infer circle, semicircle, corridor, ambush, chase line, firing line, swarm, or duel lighting := infer dust beams, muzzle flashes, backlight, sunset, neon, smoke, haze, or natural light style_blend := infer animation, photorealism, hybrid, comic, game cutscene, cinematic anime, or live-action vfx render: a dramatic confrontation scene where all elements are inferred from the inputs. forbidden: fixed anime schoolgirl, fixed agents, fixed firearms, fixed ruined building, random enemies, cluttered action, unclear protagonist.