To get this working, make sure you're using the trigger: shibari-rope-bondage
Check my images to see varying settings I used. For the Z-image variant I highly suggest inferencing with Z-Image-Turbo.
Z-Image_Shibari.safetensors