CVE-2006-3011 multiple PHP safe mode bypasses (CVE-2006-4481, CVE-2006-2563)

The error_log function in basic_functions.c in PHP before 4.4.4 and 5.x before 5.1.5 allows local users to bypass safe mode and open_basedir restrictions via a "php://" or other scheme in the third argument, which disables safe mode.

The (1) ftpd and (2) ksu programs in (a) MIT Kerberos 5 (krb5) up to 1.5, and 1.4.x before 1.4.4, and (b) Heimdal 0.7.2 and earlier, do not check return codes for setuid calls, which might allow local users to gain privileges by causing setuid to fail to drop privileges. NOTE: as of 20060808, it is not known whether an exploitable attack scenario exists for these issues.

Siemens Speedstream Wireless Router 2624 allows local users to bypass authentication and access protected files by using the Universal Plug and Play UPnP/1.0 component.

Untrusted search path vulnerability in Winlogon in Microsoft Windows 2000 SP4, when SafeDllSearchMode is disabled, allows local users to gain privileges via a malicious DLL in the UserProfile directory, aka "User Profile Elevation of Privilege Vulnerability."

Multiple unspecified vulnerabilities in IBM WebSphere Application Server before 6.1.0.1 have unspecified impact and attack vectors involving (1) "SOAP requests and responses", (2) mbean, (3) ThreadIdentitySupport, and possibly others.

Agnitum Outpost Firewall Pro 3.51.759.6511 (462), as used in (1) Lavasoft Personal Firewall 1.0.543.5722 (433) and (2) Novell BorderManager Novell Client Firewall 2.0, does not properly restrict user activities in application windows that run in a LocalSystem context, which allows local users to gain privileges and execute commands (a) via the "open folder" option when no instance of explorer.exe is running, possibly related to the ShellExecute API function; or (b) by overwriting a batch file through the "Save Configuration As" option. NOTE: this might be a vulnerability in Microsoft Windows and explorer.exe instead of the firewall.

jmx-console/HtmlAdaptor in the jmx-console in the JBoss web application server, as shipped with Cisco Security Monitoring, Analysis and Response System (CS-MARS) before 4.2.1, allows remote attackers to gain privileges as the CS-MARS administrator and execute arbitrary Java code via an invokeOp action in the BSHDeployer jboss.scripts service name.

The vCard functions in Joomla! 1.0.5 use predictable sequential IDs for vcards and do not restrict access to them, which allows remote attackers to obtain valid e-mail addresses to conduct spam attacks by modifying the contact_id parameter to index2.php.

heartbeat.c in heartbeat before 2.0.6 sets insecure permissions in a shmget call for shared memory, which allows local users to cause an unspecified denial of service via unknown vectors, possibly during a short time window on startup.

imageVue 16.1 allows remote attackers to obtain folder permission settings via a direct request to dir.php, which returns an XML document that lists folders and their permissions.

Certain XML functions in IBM DB2 8.1 run with the privileges of DB2 instead of the logged-in user, which allows remote attackers to create or overwrite files via (1) XMLFileFromVarchar or (2) XMLFileFromClob, or read files via (3) XMLVarcharFromFile or (4) XMLClobFromFile.

The Java Plug-in J2SE 1.3.0_02 through 5.0 Update 5, and Java Web Start 1.0 through 1.2 and J2SE 1.4.2 through 5.0 Update 5, allows remote attackers to exploit vulnerabilities by specifying a JRE version that contain vulnerabilities.

Joomla! before 1.0.11 does not limit access to the Admin Popups functionality, which has unknown impact and attack vectors.

Multiple unspecified vulnerabilities in Joomla! before 1.0.11, related to "Injection Flaws," allow attackers to have an unknown impact via (1) globals.php, which uses include_once() instead of require(); (2) the $options variable; (3) Admin Upload Image; (4) ->load(); (5) content submissions when frontpage is selected; (6) the mosPageNav constructor; (7) saveOrder functions; (8) the absence of "exploit blocking rules" in htaccess; and (9) the ACL.

htpasswd, as used in Acme thttpd 2.25b and possibly other products such as Apache, might allow local users to gain privileges via shell metacharacters in a command line argument, which is used in a call to the system function. NOTE: since htpasswd is normally installed as a non-setuid program, and the exploit is through command line options, perhaps this issue should not be included in CVE. However, if there are some typical or recommended configurations that use htpasswd with sudo privileges, or common products that access htpasswd remotely, then perhaps it should be included.

security flaw

Unspecified vulnerability in Adobe Flash Player before 9.0.16.0 allows user-assisted remote attackers to bypass the allowScriptAccess protection via unspecified vectors.

Microsoft Windows XP SP1 and SP2 before August 2004, and possibly other operating systems and versions, uses insecure default ACLs that allow the Authenticated Users group to gain privileges by modifying critical configuration information for the (1) Simple Service Discovery Protocol (SSDP), (2) Universal Plug and Play Device Host (UPnP), (3) NetBT, (4) SCardSvr, (5) DHCP, and (6) DnsCache services, aka "Permissive Windows Services DACLs." NOTE: the NetBT, SCardSvr, DHCP, DnsCache already require privileged access to exploit.

fantastico in Cpanel does not properly handle when it has insufficient permissions to perform certain file operations, which allows remote authenticated users to obtain the full pathname, which is leaked in a PHP error message.

ISNTSmtp directory in Trend Micro InterScan Messaging Security Suite (IMSS) 5.5 build 1183 and possibly other versions before 5.7.0.1121, uses insecure DACLs for critical files, which allows local users to gain SYSTEM privileges by modifying ISNTSysMonitor.exe.

madvise_remove in Linux kernel 2.6.16 up to 2.6.16.6 does not follow file and mmap restrictions, which allows local users to bypass IPC permissions and replace portions of readonly tmpfs files with zeroes, aka the MADV_REMOVE vulnerability. NOTE: this description was originally written in a way that combined two separate issues. The mprotect issue now has a separate name, CVE-2006-2071.