Pontus Skoeld <[EMAIL PROTECTED]> writes: > Hi, > > in addition to Niels comments, here are some unstructured thoughts - I > do realize that this is work in progress, but I want to point it out > so I don't forget. Also consider this food for thought, I may be wrong > about things.
Hello, and thanks for the suggestions. If you could look at the code again, when I believe I'm satisfied with it, it would be even better. :-) > I assume there will be a check if the given mechanism is available > here? Yup. > If I'm not missing something, the current implementation would allow > an attacker to request authentication as user foo and use GSS to > authenticate herself as bar, and gain access as foo, which would be a > bad thing Yup. > I assume GSS can be told to only accept authentication for a given > user or that one can check after authentication who was > authenticated? You can extract the identity of the user who was authenticated, and then carry out the authorization procedure. Is there an authorization infrastructure in lsh? If not, one should probably simply try to log in as the authenticated GSSAPI identity instead of the requested username (the username can be empty according to the GSS-SSH specification). Thanks, Simon _______________________________________________ lsh-bugs mailing list [EMAIL PROTECTED] http://lists.lysator.liu.se/mailman/listinfo/lsh-bugs
