security flaw

The prescan function in Sendmail 8.12.9 allows remote attackers to execute arbitrary code via buffer overflow attacks, as demonstrated using the parseaddr function in parseaddr.c.

Buffer overflow in CoreFoundation in Mac OS X 10.3.9 allows attackers to execute arbitrary code via command line arguments to an application that uses CoreFoundation.

The System Profiler in Mac OS X 10.4.2 labels a Bluetooth device with "Requires Authentication: No" even when the user has selected the "Require pairing for security" option, which could confuse users about which setting is valid.

Unknown vulnerability in LoginWindow for Mac OS X 10.3.4, related to "handling of directory services lookups."

security flaw

Multiple buffer overflows in Cyrus SASL library 2.1.9 and earlier allow remote attackers to cause a denial of service and possibly execute arbitrary code via (1) long inputs during user name canonicalization, (2) characters that need to be escaped during LDAP authentication using saslauthd, or (3) an off-by-one error in the log writer, which does not allocate space for the null character that terminates a string.

AppKit for Mac OS X 10.3.9 and 10.4.2 allows attackers with physical access to create local accounts by forcing a particular error to occur at the login window.

Buffer overflow in AppKit for Mac OS X 10.3.9 and 10.4.2, as used in applications such as TextEdit, allows external user-assisted attackers to execute arbitrary code via a crafted Microsoft Word file.

Unspecified vulnerability in Mac OS X before 10.3.4 has unknown impact and attack vectors related to "logging when tracing system calls."

Buffer overflow in AppKit for Mac OS X 10.3.9 and 10.4.2 allows external user-assisted attackers to execute arbitrary code via a crafted Rich Text Format (RTF) file.

Unspecified vulnerability in the Apple Mac OS X kernel before 10.4.2 allows remote attackers to cause a denial of service (kernel panic) via a crafted TCP packet, possibly related to source routing or loose source routing.

Argument injection vulnerability in the SSH URI handler for Safari on Mac OS 10.3.3 and earlier allows remote attackers to (1) execute arbitrary code via the ProxyCommand option or (2) conduct port forwarding via the -R option.

security flaw

A "potential buffer overflow in ruleset parsing" for Sendmail 8.12.9, when using the nonstandard rulesets (1) recipient (2), final, or (3) mailer-specific envelope recipients, has unknown consequences.

security flaw

OpenSSL 0.9.6d and earlier, and 0.9.7-beta2 and earlier, does not properly handle ASCII representations of integers on 64 bit platforms, which could allow attackers to cause a denial of service and possibly execute arbitrary code.

Dashboard in Apple Mac OS X Tiger 10.4 allows attackers to execute arbitrary commands by overriding the behavior of system widgets via a user widget with the same bundle identifier (CFBundleIdentifier), a different vulnerability than CVE-2005-1474.

MCX Client for Apple Mac OS X 10.4.x up to 10.4.1 insecurely logs Portable Home Directory credentials, which allows local users to obtain the credentials.

HelpViewer in Mac OS X 10.3.3 and 10.2.8 processes scripts that it did not initiate, which can allow attackers to execute arbitrary code, an issue that was originally reported as a directory traversal vulnerability in the Safari web browser using the runscript parameter in a help: URI handler.

Apple Mac OS X 10.4.x up to 10.4.1 sets insecure world- and group-writable permissions for the (1) system cache folder and (2) Dashboard system widgets, which allows local users to conduct unauthorized file operations via "file race conditions."

The CoreGraphics Window Server in Mac OS X 10.4.1 allows local users with console access to gain privileges by "launching commands into root sessions."

The default protocol helper for the disk: URI on Mac OS X 10.3.3 and 10.2.8 allows remote attackers to write arbitrary files by causing a disk image file (.dmg) to be mounted as a disk volume.

Workgroup Manager in Apple Mac OS X Server 10.2 through 10.2.6 does not disable a password for a new account before it is saved for the first time, which allows remote attackers to gain unauthorized access via the new account before it is saved.