{
  «task»: «high-end photorealistic composite image generation»,
  «model_intent»: «reference-driven scene reconstruction»,
  «references»: {
    «reference_1»: {
      «role»: «primary subject identity»,
      «description»: «Use this image as the sole identity source. Extract facial structure, body proportions, skin texture, hair, and overall likeness from this reference. Do not stylize or alter identity.»
    },
    «reference_2»: {
      «role»: «screen UI layout reference»,
      «description»: «Use this image strictly as the phone screen content layout. Recreate the UI composition, spacing, typography scale, and interface proportions exactly as seen.»
    }
  },
  «scene_description»: {
    «environment»: «minimal studio background with soft neutral tones»,
    «main_object»: «iPhone 17 Pro placed flat in a realistic perspective, slightly angled, premium metallic edges visible»,
    «screen_state»: «phone screen is ON and active»,
    «composition»: «the primary subject stands physically on top of the phone screen, feet naturally grounded on the glass surface»
  },
  «subject_handling»: {
    «identity_lock»: true,
    «pose»: «natural relaxed standing pose, confident casual posture»,
    «scale»: «subject proportionally scaled to appear realistically standing on a smartphone screen»,
    «integration»: «accurate contact shadows between shoes and glass surface, realistic weight distribution»
  },
  «wardrobe_styling»: {
    «outfit_source»: «Kith official website»,
    «style»: «modern premium streetwear»,
    «outfit_description»: «a realistic Kith outfit combination including top, bottom, footwear and accessories, matching current Kith lookbook aesthetics»,
    «fabric_detail»: «high-resolution textile texture, natural folds, correct fabric weight and stitching»,
    «branding»: «subtle and realistic Kith branding, no exaggeration»
  },
  «phone_screen_content»: {
    «ui_type»: «Instagram post view»,
    «content_mapping»: «place the second reference image exactly onto the phone screen as if Instagram is open»,
    «ui_realism»: «accurate glass reflections, screen glare, brightness falloff, and OLED contrast»,
    «perspective_warp»: «screen content follows the physical curvature and perspective of the phone»
  },
  «lighting»: {
    «key_light»: «soft studio key light from upper left»,
    «fill_light»: «gentle fill to preserve shadow detail»,
    «reflections»: «realistic reflections on phone glass and metallic edges»,
    «shadow_behavior»: «physically correct shadows cast by subject onto the phone screen»
  },
  «camera»: {
    «angle»: «slightly elevated three-quarter angle»,
    «lens»: «50mm equivalent»,
    «depth_of_field»: «moderate DOF, subject and phone both sharp»,
    «realism»: «photographic realism, no illustration or CGI look»
  },
  «quality_constraints»: {
    «photorealism»: «ultra realistic»,
    «resolution»: «high resolution»,
    «no_artifacts»: true,
    «no_cartoon»: true,
    «no_anime»: true,
    «no_surreal_distortion»: true
  },
  «final_instruction»: «The final image must look like a real premium product campaign photo, as if shot in a professional studio for a high-end fashion and tech brand collaboration.»
}

Sustituye la captura de pantalla que aparece en el iPhone de @image1 por @image2 sin cambiar nada más.

the man on top of the phone looks around in a cool way and stands naturally, the image on the phone screen scrolls upward and a new image appears