Phone Stuck to the Train AI Prompt: Swap Yourself In

[Edit Goal]
Edit @Video1. Across the entire video, replace only the young man in the black double-breasted blazer with the woman from @Image1. She performs every one of his movements exactly. Keep the original subway station and train.
[Source Video Role]
@Video1 is the sole editing master. It defines the subway station, the silver train, the passengers inside, the background commuters, the composition, the camera position and movement, the cut at about 2.7 seconds, the motion path, body poses, gestures, mouth movements, timing, and event order.
[How the Video Works]
Shot 1 (0 to 2.7 s), filmed by a friend on the platform: she runs along the yellow platform edge to the stopped silver train, presses her phone flat against the outside of a train window with the screen facing the glass and the camera facing the platform, and sticks it there. She taps and smooths it with both hands so it holds to the glass, while the passengers inside watch.
Shot 2 (2.7 s to end) is filmed by that phone, now stuck on the outside of the train window and looking out at the platform. The train starts moving, so the camera glides sideways along the platform with the train. She stays close to the phone and walks, then runs alongside the moving train to keep her face in front of it, mouthing the song straight into the camera, pulling at her blazer collar, crouching, lunging, and swinging her arms. The train accelerates, she cannot keep up, she falls behind and drops out of frame, and the empty platform keeps sliding past.
[Target Reference Role]
@Image1 is a character sheet of one single woman: a close-up of her face plus front, left profile, back, and right profile views of the same person. Use only her face, skin with light freckles, green eyes, dark-red-to-copper wavy hair in a high messy ponytail with loose face-framing strands, and her outfit: oversized lime-green double-breasted linen blazer with dark buttons and rolled sleeves showing brown cuffs, a dark-brown collared shirt, wide-leg lime-green trousers, and pointed dark-brown shoes. Do not use the light-grey studio background, the white studio lighting, the static standing poses, or the multi-panel layout.
[Edit Scope]
Modify only the man’s body and the area his silhouette covers. The entire video contains exactly one woman from @Image1 and no trace of the man. Remove his white earphone cable, white shirt, black tie, and blue jeans; she wears only her own outfit. Do not modify the train, the window, the platform, the signs, the passengers, the commuters, or the camera.
[Timeline Inheritance]
She inherits every appearance, motion, and exit of the man, including timing, duration, path, and speed changes. Her ponytail and the loose blazer swing with every move. Her face stays recognisably the woman from @Image1 in every frame, including fast motion and motion blur. Keep the real smartphone footage look of @Video1: the same subway light on her skin and hair, the same grain, and the same blur.
Except for the subject replaced above, keep all other people, props, scene content, camera movements, cuts, and event order from @Video1 unchanged.
[Lip Sync]
Her lips copy the man’s mouth movements exactly, word for word and syllable for syllable, with the same timing: every opening, closing, rounded vowel, pause, and big open-mouth note lands on the same frame as in @Video1. Do not invent new words, new mouth shapes, or a different melody.
[Audio]
Only natural subway ambience: <footsteps on the platform>, <train doors closing>, <train rumbling as it pulls away>, <muffled crowd chatter>. No music, no singing, no voice.