The scenario looks at each new image, picks tags from its content, and writes keywords right into the file.