The code makes a call to the URL you noted, sending along the user's IP, your site's domain, and the user's useragent string. It's then printing onto your site any code it receives from the cURL request. The code received could be anything. It could be HTML, JavaScript, or any other client side code. It's probably not server-side code since there's no eval()
running the code received.
It appears to target Internet Explorer, Chrome, and FireFox browsers, but not crawlers/bots.
EDIT: As FDL pointed out in his comment, this appears to be printing only if it receives a string where the second, third, and fourth characters are scr
, meaning it likely only prints to the page if it received a <script>
tag.