Google Chrome before 1.0.154.46 does not properly restrict access from web pages to the (1) Set-Cookie and (2) Set-Cookie2 HTTP response headers, which allows remote attackers to obtain sensitive information from cookies via XMLHttpRequest calls and other web script.

DISPUTED Google Chrome 1.0.154.43 allows remote attackers to trick a user into visiting an arbitrary URL via an onclick action that moves a crafted element to the current mouse position, related to a "Clickjacking" vulnerability. NOTE: a third party disputes the relevance of this issue, stating that "every sufficiently featured browser is and likely will remain susceptible to the behavior known as clickjacking," and adding that the exploit code "is not a valid demonstration of the issue."

Cross-domain vulnerability in the V8 JavaScript engine in Google Chrome before 1.0.154.46 allows remote attackers to bypass the Same Origin Policy via a crafted script that accesses another frame and reads its full URL and possibly other sensitive information, or modifies the URL of this frame.

Integer overflow in the CrossCallParamsEx::CreateFromBuffer function in sandbox/src/crosscallserver.cc in Google Chrome before 4.0.249.89 allows attackers to leverage renderer access to cause a denial of service (heap memory corruption) or possibly have unspecified other impact via a malformed message, related to deserializing of sandbox messages.

Google Chrome before 4.0.249.78 sends an https URL in the Referer header of an http request in certain circumstances involving https to http redirection, which allows remote HTTP servers to obtain potentially sensitive information via standard HTTP logging.

The image decoder in WebKit before r52833, as used in Google Chrome before 4.0.249.78, does not properly handle a failure of memory allocation, which allows remote attackers to execute arbitrary code in the Chrome sandbox via a malformed GIF file that specifies a large size.

WebKit before r51295, as used in Google Chrome before 4.0.249.78, presents a directory-listing page in response to an XMLHttpRequest for a file:/// URL that corresponds to a directory, which allows attackers to obtain sensitive information or possibly have unspecified other impact via a crafted local HTML document.

Google Chrome before 4.0.249.89, when a SOCKS 5 proxy server is configured, sends DNS queries directly, which allows remote DNS servers to obtain potentially sensitive information about the identity of a client user via request logging, as demonstrated by a proxy server that was configured for the purpose of anonymity.

Google Chrome before 4.0.249.89 attempts to make direct connections to web sites when all configured proxy servers are unavailable, which allows remote HTTP servers to obtain potentially sensitive information about the identity of a client user via standard HTTP logging, as demonstrated by a proxy server that was configured for the purpose of anonymity.

The ParamTraits<SkBitmap>::Read function in common/commonparamtraits.cc in Google Chrome before 4.0.249.78 does not use the correct variables in calculations designed to prevent integer overflows, which allows attackers to leverage renderer access to cause a denial of service or possibly have unspecified other impact via bitmap data, related to deserialization.

WebCore/bindings/v8/custom/V8DOMWindowCustom.cpp in WebKit before r52401, as used in Google Chrome before 4.0.249.78, allows remote attackers to bypass the Same Origin Policy via vectors involving the window.open method.

Use-after-free vulnerability in Google Chrome before 4.0.249.78 allows user-assisted remote attackers to cause a denial of service (application crash) or possibly execute arbitrary code via vectors involving the display of a blocked popup window during navigation to a different web site.

WebKit before r53525, as used in Google Chrome before 4.0.249.89, allows remote attackers to execute arbitrary code in the Chrome sandbox via a malformed RUBY element, as demonstrated by a <ruby>><table><rt> sequence.

Multiple integer overflows in factory.cc in Google V8 before r3560, as used in Google Chrome before 4.0.249.89, allow remote attackers to execute arbitrary code in the Chrome sandbox via crafted use of JavaScript arrays.

The ParamTraits<SkBitmap>::Read function in common/commonparamtraits.cc in Google Chrome before 4.0.249.78 does not initialize the memory locations that will hold bitmap data, which might allow remote attackers to obtain potentially sensitive information from process memory by providing insufficient data, related to use of a (1) thumbnail database or (2) HTML canvas.

Multiple integer signedness errors in factory.cc in Google V8 before r3560, as used in Google Chrome before 4.0.249.89, allow remote attackers to execute arbitrary code in the Chrome sandbox via crafted use of JavaScript arrays.

Stack consumption vulnerability in the ChildProcessSecurityPolicy::CanRequestURL function in browser/childprocesssecuritypolicy.cc in Google Chrome before 4.0.249.78 allows remote attackers to cause a denial of service (memory consumption and application crash) via a URL that specifies multiple protocols, as demonstrated by a URL that begins with many repetitions of the view-source: substring.

Multiple integer overflows in Skia, as used in Google Chrome before 4.0.249.78, allow remote attackers to execute arbitrary code in the Chrome sandbox or cause a denial of service (memory corruption and application crash) via vectors involving CANVAS elements.

Google Chrome before 4.0.249.78 on Windows does not perform the expected encoding, escaping, and quoting for the URL in the --app argument in a desktop shortcut, which allows user-assisted remote attackers to execute arbitrary programs or obtain sensitive information by tricking a user into creating a crafted shortcut.

browser/login/loginprompt.cc in Google Chrome before 4.0.249.89 populates an authentication dialog with credentials that were stored by Password Manager for a different web site, which allows user-assisted remote HTTP servers to obtain sensitive information via a URL that requires authentication, as demonstrated by a URL in the SRC attribute of an IMG element.

The Google URL Parsing Library (aka google-url or GURL) in Google Chrome before 4.1.249.1064 allows remote attackers to bypass the Same Origin Policy via unspecified vectors.

Google Chrome before 4.1.249.1064 does not properly handle fonts, which allows remote attackers to cause a denial of service (memory corruption) and possibly have unspecified other impact via unknown vectors.

Google Chrome before 4.1.249.1036 processes HTTP headers before invoking the SafeBrowsing feature, which allows remote attackers to have an unspecified impact via crafted headers.

Multiple race conditions in the sandbox infrastructure in Google Chrome before 4.1.249.1036 have unspecified impact and attack vectors.

Unspecified vulnerability in Google Chrome before 4.1.249.1036 allows remote attackers to trigger the omission of a download warning dialog via unknown vectors.

Google Chrome before 4.1.249.1036 allows remote attackers to cause a denial of service (memory error) or possibly have unspecified other impact via a malformed SVG document.

The sandbox infrastructure in Google Chrome before 4.1.249.1036 does not properly use pointers, which has unspecified impact and attack vectors.

Buffer overflow in the browser kernel in Google Chrome before 2.0.172.33 allows remote HTTP servers to cause a denial of service (application crash) or possibly execute arbitrary code via a crafted response.

Google Chrome before 1.0.154.53 displays a cached certificate for a (1) 4xx or (2) 5xx CONNECT response page returned by a proxy server, which allows man-in-the-middle attackers to spoof an arbitrary https site by letting a browser obtain a valid certificate from this site during one request, and then sending the browser a crafted 502 response page upon a subsequent request.

src/net/http/httptransactionwinhttp.cc in Google Chrome before 1.0.154.53 uses the HTTP Host header to determine the context of a document provided in a (1) 4xx or (2) 5xx CONNECT response from a proxy server, which allows man-in-the-middle attackers to execute arbitrary web script by modifying this CONNECT response, aka an "SSL tampering" attack.