MAX — IMAGE-TO-VIDEO SOURCE PACK v1

Six new pose/angle renders plus the two approved identity references.
All files are full-resolution PNGs. Generated concept imagery, not built hardware.

Identity: Pearl ceramic armour, graphite joints, narrow blue chest indicator,
rounded black-glass visor, two horizontal sensor bars and circular side discs.
Preserve panel geometry and proportions. No lips, nose, eyebrows or human eyeballs.
The rear view is a design extrapolation; final engineering is unannounced.

Suggested image-to-video prompts:
Front / three-quarter: locked camera, subtle head turn, tiny weight shift; 4–6 seconds.
Profile: slow 15-degree camera orbit, feet fixed, preserve silhouette; 4–6 seconds.
Rear: slow rear orbit, no transformation or new equipment; 4–6 seconds.
Walking: one measured step, natural opposing arm swing, no sliding feet; 3–5 seconds.
Protective: gently lift/settle open palm, calm stance, no weapons or combat; 4–6 seconds.
Hero: open hand gesture with a small head turn; camera slowly pushes in; 4–6 seconds.
Portrait: minute head rotation, two fixed sensor bars, no blinking human eyes.

Use the approved hero and portrait as identity references alongside the starting frame.
Ask your video model to avoid armour morphing, extra digits, floating feet, camera
shake, facial features and added labels. Keep videos silent and export H.264 MP4.
These are independent still generations, not calibrated multiview geometry or exact
consecutive animation frames. Check identity continuity when joining clips.