rendering/render_replaced.cpp in Konqueror in KDE before 4.9.3 allows remote attackers to cause a denial of service (NULL pointer dereference) via a crafted web page, related to "trying to reuse a frame with a null part."

kdesktop_lock in kdebase before 3.1.3-5.11 for KDE in Red Hat Enterprise Linux (RHEL) 3 does not properly terminate, which can prevent the screensaver from activating or prevent users from manually locking the desktop.

langen2kvtml in KDE 3.0 to 3.4.2 creates insecure temporary files in /tmp with predictable names, which allows local users to overwrite arbitrary files.

The (1) Kate and (2) Kwrite applications in KDE KDE 3.2.x through 3.4.0 do not properly set the same permissions on the backup file as were set on the original file, which could allow local users and possibly remote attackers to obtain sensitive information.

KPPP 2.1.2 in KDE 3.1.5 and earlier, when setuid root without certain wrappers, does not properly close a privileged file descriptor for a domain socket, which allows local users to read and write to /etc/hosts and /etc/resolv.conf and gain control over DNS name resolution by opening a number of file descriptors before executing kppp.

The patch for integer overflow vulnerabilities in Xpdf 2.0 and 3.0 (CVE-2004-0888) is incomplete for 64-bit architectures on certain Linux distributions such as Red Hat, which could leave Xpdf users exposed to the original vulnerabilities.

Kommander in KDE 3.2 through KDE 3.4.0 executes data files without confirmation from the user, which allows remote attackers to execute arbitrary code.

Konqueror in KDE 3.2.3 and earlier allows web sites to set cookies for country-specific top-level domains, such as .ltd.uk, .plc.uk and .firm.in, which could allow remote attackers to perform a session fixation attack and hijack a user's HTTP session.

Buffer overflow in the VCF file information reader for KDE Personal Information Management (kdepim) suite in KDE 3.1.0 through 3.1.4 allows attackers to execute arbitrary code via a VCF file.

KDM in KDE 3.1.3 and earlier does not verify whether the pam_setcred function call succeeds, which may allow attackers to gain root privileges by triggering error conditions within PAM modules, as demonstrated in certain configurations of the MIT pam_krb5 module.