add static/robots.txt, disabling crawling of /api/
This is related to T662, but doesn't completely addresses it as we also need an HTTP-level alias /robots.txt -> /static/robots.txt.