You could strongly argue that collecting clickstream and other user browser session info via a toolbar is not a form of web robot (crawler, spider, etc.), and thus robots.txt does not apply.