FreeBSD port programs that use libkvm for FreeBSD 4.6.2-RELEASE and earlier, including (1) asmon, (2) ascpu, (3) bubblemon, (4) wmmon, and (5) wmnet2, leave open file descriptors for /dev/mem and /dev/kmem, which allows local users to read kernel memory.

Multiple buffer overflows in purity 1-16 allow local users to gain privileges and modify high scores tables.

Buffer overflow in the authentication function for Microsoft SQL Server 2000 and Microsoft Desktop Engine (MSDE) 2000 allows remote attackers to execute arbitrary code via a long request to TCP port 1433, aka the "Hello" overflow.

Buffer overflow in the parsing mechanism for ISS Internet Scanner 6.2.1, when using the license banner HTTP check, allows remote attackers to execute arbitrary code via a long web server response.

SMTP content filter engines, including (1) GFI MailSecurity for Exchange/SMTP before 7.2, (2) InterScan VirusWall before 3.52 build 1494, (3) the default configuration of MIMEDefang before 2.21, and possibly other products, do not detect fragmented emails as defined in RFC2046 ("Message Fragmentation and Reassembly") and supported in such products as Outlook Express, which allows remote attackers to bypass content filtering, including virus checking, via fragmented emails of the message/partial content type.

Buffer overflow in Savant Web Server 3.1 and earlier allows remote attackers to execute arbitrary code via a long HTTP GET request.

os._execvpe from os.py in Python 2.2.1 and earlier creates temporary files with predictable names, which could allow local users to execute arbitrary code via a symlink attack.

TNS Listener in Oracle Net Services for Oracle 9i 9.2.x and 9.0.x, and Oracle 8i 8.1.x, allows remote attackers to cause a denial of service (hang or crash) via a SERVICE_CURLOAD command.

Veritas Backup Exec 8.5 and earlier requires that the "RestrictAnonymous" registry key for Microsoft Exchange 2000 must be set to 0, which enables anonymous listing of the SAM database and shares.

The "View Bugs" page (view_all_bug_page.php) in Mantis 0.17.4a and earlier includes summaries of private bugs for users that do not have access to any projects.

Mantis 0.17.4a and earlier allows remote attackers to view private bugs by modifying the f_id bug ID parameter to (1) bug_update_advanced_page.php, (2) bug_update_page.php, (3) view_bug_advanced_page.php, or (4) view_bug_page.php.

config_inc2.php in Mantis before 0.17.4 allows remote attackers to execute arbitrary code or read arbitrary files via the parameters (1) g_bottom_include_page, (2) g_top_include_page, (3) g_css_include_file, (4) g_meta_include_file, or (5) a cookie.

summary_graph_functions.php in Mantis 0.17.3 and earlier allows remote attackers to execute arbitrary PHP code by modifying the g_jpgraph_path parameter to reference the location of the PHP code.

Mantis before 0.17.4 allows remote attackers to list project bugs without authentication by modifying the cookie that is used by the "View Bugs" page.

print_all_bug_page.php in Mantis 0.17.3 and earlier does not verify the limit_reporters option, which allows remote attackers to view bug summaries for bugs that would otherwise be restricted.

Multiple SQL injection vulnerabilities in Mantis 0.17.2 and earlier, when running without magic_quotes_gpc enabled, allows remote attackers to gain privileges or perform unauthorized database operations via modified form fields, e.g. to account_update.php.

securetar, as used in AMaViS shell script 0.2.1 and earlier, allows users to cause a denial of service (CPU consumption) via a malformed TAR file, possibly via an incorrect file size parameter.

Cisco Virtual Private Network (VPN) Client software 2.x.x, and 3.x before 3.6(Rel), when configured with all tunnel mode, can be forced into acknowledging a TCP packet from outside the tunnel.

Cisco Virtual Private Network (VPN) Client software 2.x.x, and 3.x before 3.5.2B, does not generate sufficiently random numbers, which may make it vulnerable to certain attacks such as spoofing.

Cisco Virtual Private Network (VPN) Client software 2.x.x, and 3.x before 3.5.1C, does not properly verify that certificate DN fields match those of the certificate from the VPN Concentrator, which allows remote attackers to conduct man-in-the-middle attacks.