Il 23/05/2013 20:43, Peter Lebbing ha scritto: >> Really useful, IMVHO. Unless you have to sign *a lot* of things... > Werner Koch does not agree it's a security feature (and I suppose that's why > you > think it's useful), as he said in this[1] thread: > [1] http://lists.gnupg.org/pipermail/gnupg-users/2013-February/046051.html
Similar threads appeared on OpenSC ML too. That's why I was investigating a "port" of OpenPGPCarf to Yubico token (that offers a button that can be read by the Java code -- too bad it requires a library available from NXP only under strict NDA :( A less robust (against invasive attacks) option could be the GNUK token. >> In any case it is not a security measure because the host may simply >> cache the PIN and and silently do a verify command before each sign >> operation. To avoid that simple workaround, a pinpad reader which >> filters the VERIFY command would be needed. The host may cache it only if it ever sees it :) There exists cards with button and display: having an OOB bidirectional channel can give much more security... Another option could be a HOTP code instead of a static PIN (maybe I'll include this in MyPGPid :) ). BYtE, Diego. _______________________________________________ Gnupg-users mailing list [email protected] http://lists.gnupg.org/mailman/listinfo/gnupg-users
