Postnuke XSS issues [correction]

From: Mark Grimes (markat_private)
Date: Thu Sep 26 2002 - 09:09:08 PDT

  • Next message: Jean-loup Gailly: "remote SYSTEM compromise in WASD OpenVMS http server"

    As it turns out the Postnuke issue in particular is a red herring.
    
    As the lead developer describes it -- the cookie generated is a local
    site cookie that is sandboxed within the confines of the
    browser/session.
    
    It is not the remote user's cookie.
    
    It is easy to be fooled by such a vulnerability if the local site cookie
    is empty as well as the remote user's cookie.  Some conditions can
    generate the exact same look and feel.
    
    Be warned that all instances of scriptable java within URL/HTML
    constructs (even with document.cookie) may be not really be an XSS
    issue even if it walks talks and acts like an XSS bug.
    
    Only carrying out the full exploit (cookie theft/account hijack would
    prove if it is really an issue in these cases.  However, I chose the
    alternative and obtained feedback from the author.
    
    [The feedback came much later then post to bugtraq, there was such a
    long delay I thought the post was moderated -- since it did get posted,
    this message serves as a correction.]
    
    -- 
    Mark Grimes <markat_private>
    Stateful Labs
    



    This archive was generated by hypermail 2b30 : Thu Sep 26 2002 - 14:05:24 PDT