For images marked with the class .include-dark, the extension includes two images:

1. The original image with the class .light-content
2. A dark version of the image, assumed to have the same filename with an additional suffix -dark, with the class .dark-content

For example this markdown:

![](image.png){.border .include-dark fig-alt="A probably long description of this figure"}

Results in output equivalent to this markdown:

![](image.png){.border .light-content fig-alt="A probably long description of this figure"}
![](image-dark.png){.border .dark-content fig-alt="A probably long description of this figure"}