Stephan Wehner wrote:
I think positive action should be taken, not negative action
(Highlighting for known domains is a negative action when you're
concerned with unknown possibly spoofed domains.)
Much brainstorming needed! A few comments
below.
1. Set up a "protected list". This is a list of domains which may
possibly be interesting for an attack. A protected list may be derived
from the history information as mentioned, or, why not charge
interested parties to be included in the protected list.
That's essentially what the CA signed cert
is! An interested party goes and gets on
a "list" ... it is a "minor technical detail" that
this list is distributed by means of certs, and
in the future the list is going to be distributed
by means of callbacks (OCSP).
2. Compare the image of the text as it appears on the screen to the
image as they would appear for the items in the "protected list", and
mark it /alert if it is very similar. (Positive marking instead of
negative marking)
What could be wrong with that?
Bear in mind that your attacker has access
to the code that would compare these two
images....
But instead of highlighting / alert popup, how about switching to a
different theme, which is a more general way of putting it: the
familiar web site theme, vs. the possibly dangerous website theme, vs.
the known to be evil website theme, vs.etc etc.?
Yes, this is what TrustBar.mozdev tries to do at
some level. I think the display of "I don't know
where this is at!" is a valuable thing from the
browser.
iang
--
News and views on what matters in finance+crypto:
http://financialcryptography.com/
_______________________________________________
Mozilla-security mailing list
[email protected]
http://mail.mozilla.org/listinfo/mozilla-security
