The parse function in llhttp < 2.1.4 and < 6.0.6. ignores chunk extensions when parsing the body of chunked requests. This leads to HTTP Request Smuggling (HRS) under certain conditions.

In Apache Tomcat 9.0.0.M1 to 9.0.30, 8.5.0 to 8.5.50 and 7.0.0 to 7.0.99 the HTTP header parsing code used an approach to end-of-line parsing that allowed some invalid HTTP headers to be parsed as valid. This led to a possibility of HTTP Request Smuggling if Tomcat was located behind a reverse proxy that incorrectly handled the invalid Transfer-Encoding header in a particular manner. Such a reverse proxy is considered unlikely.

The refactoring present in Apache Tomcat 9.0.28 to 9.0.30, 8.5.48 to 8.5.50 and 7.0.98 to 7.0.99 introduced a regression. The result of the regression was that invalid Transfer-Encoding headers were incorrectly processed leading to a possibility of HTTP Request Smuggling if Tomcat was located behind a reverse proxy that incorrectly handled the invalid Transfer-Encoding header in a particular manner. Such a reverse proxy is considered unlikely.

The CGIHandler class in Python before 2.7.12 does not protect against the HTTP_PROXY variable name clash in a CGI script, which could allow a remote attacker to redirect HTTP requests.

A timing side-channel in the handling of RSA ClientKeyExchange messages was discovered in GnuTLS. This side-channel can be sufficient to recover the key encrypted in the RSA ciphertext across a network in a Bleichenbacher style attack. To achieve a successful decryption the attacker would need to send a large amount of specially crafted messages to the vulnerable server. By recovering the secret from the ClientKeyExchange message, the attacker would be able to decrypt the application data exchanged over that connection.

mod_auth_openidc is an authentication/authorization module for the Apache 2.x HTTP server that functions as an OpenID Connect Relying Party, authenticating users against an OpenID Connect Provider. In versions prior to 2.4.9.4, the 3rd-party init SSO functionality of mod_auth_openidc was reported to be vulnerable to an open redirect attack by supplying a crafted URL in the `target_link_uri` parameter. A patch in version 2.4.9.4 made it so that the `OIDCRedirectURLsAllowed` setting must be applied to the `target_link_uri` parameter. There are no known workarounds aside from upgrading to a patched version.

Insufficient validation of untrusted input in Intents in Google Chrome on Android prior to 95.0.4638.69 allowed a remote attacker to arbitrarily browser to a malicious URL via a crafted HTML page.

Swift3 before 1.9 allows remote attackers to conduct replay attacks via an Authorization request that lacks a Date header.

textview_uri_security_check in textview.c in Claws Mail before 3.18.0, and Sylpheed through 3.7.0, does not have sufficient link checks before accepting a click.

FastAPI is a web framework for building APIs with Python 3.6+ based on standard Python type hints. FastAPI versions lower than 0.65.2 that used cookies for authentication in path operations that received JSON payloads sent by browsers were vulnerable to a Cross-Site Request Forgery (CSRF) attack. In versions lower than 0.65.2, FastAPI would try to read the request payload as JSON even if the content-type header sent was not set to application/json or a compatible JSON media type (e.g. application/geo+json). A request with a content type of text/plain containing JSON data would be accepted and the JSON data would be extracted. Requests with content type text/plain are exempt from CORS preflights, for being considered Simple requests. The browser will execute them right away including cookies, and the text content could be a JSON string that would be parsed and accepted by the FastAPI application. This is fixed in FastAPI 0.65.2. The request data is now parsed as JSON only if the content-type header is application/json or another JSON compatible media type like application/geo+json. It's best to upgrade to the latest FastAPI, but if updating is not possible then a middleware or a dependency that checks the content-type header and aborts the request if it is not application/json or another JSON compatible content type can act as a mitigating workaround.

mod_auth_openidc is an authentication/authorization module for the Apache 2.x HTTP server that functions as an OpenID Connect Relying Party, authenticating users against an OpenID Connect Provider. In versions prior to 2.4.9, `oidc_validate_redirect_url()` does not parse URLs the same way as most browsers do. As a result, this function can be bypassed and leads to an Open Redirect vulnerability in the logout functionality. This bug has been fixed in version 2.4.9 by replacing any backslash of the URL to redirect with slashes to address a particular breaking change between the different specifications (RFC2396 / RFC3986 and WHATWG). As a workaround, this vulnerability can be mitigated by configuring `mod_auth_openidc` to only allow redirection whose destination matches a given regular expression.

An open redirect was discovered in Symfony 2.7.x before 2.7.50, 2.8.x before 2.8.49, 3.x before 3.4.20, 4.0.x before 4.0.15, 4.1.x before 4.1.9 and 4.2.x before 4.2.1. By using backslashes in the `_failure_path` input field of login forms, an attacker can work around the redirection target restrictions and effectively redirect the user to any domain after login.

Open redirect vulnerability in the Context UI module in the Context module 7.x-3.x before 7.x-3.6 for Drupal allows remote attackers to redirect users to arbitrary web sites and conduct phishing attacks via a URL in the destination parameter.

Node.js versions which bundle an unpatched version of OpenSSL or run against a dynamically linked version of OpenSSL which are unpatched are vulnerable to the Marvin Attack - https://people.redhat.com/~hkario/marvin/, if PCKS #1 v1.5 padding is allowed when performing RSA descryption using a private key.

macaron before 1.3.7 has an open redirect in the static handler, as demonstrated by the http://127.0.0.1:4000//example.com/ URL.

The Host Authorization middleware in Action Pack before 6.1.2.1, 6.0.3.5 suffers from an open redirect vulnerability. Specially crafted `Host` headers in combination with certain "allowed host" formats can cause the Host Authorization middleware in Action Pack to redirect users to a malicious website. Impacted applications will have allowed hosts with a leading dot. When an allowed host contains a leading dot, a specially crafted `Host` header can be used to redirect to a malicious website.

The core.urlresolvers.reverse function in Django before 1.4.14, 1.5.x before 1.5.9, 1.6.x before 1.6.6, and 1.7 before release candidate 3 does not properly validate URLs, which allows remote attackers to conduct phishing attacks via a // (slash slash) in a URL, which triggers a scheme-relative URL to be generated.

OpenSSL before 0.9.8za, 1.0.0 before 1.0.0m, and 1.0.1 before 1.0.1h does not properly restrict processing of ChangeCipherSpec messages, which allows man-in-the-middle attackers to trigger use of a zero-length master key in certain OpenSSL-to-OpenSSL communications, and consequently hijack sessions or obtain sensitive information, via a crafted TLS handshake, aka the "CCS Injection" vulnerability.

ntpkeygen can generate keys that ntpd fails to parse. NTPsec 1.2.0 allows ntpkeygen to generate keys with '#' characters. ntpd then either pads, shortens the key, or fails to load these keys entirely, depending on the key type and the placement of the '#'. This results in the administrator not being able to use the keys as expected or the keys are shorter than expected and easier to brute-force, possibly resulting in MITM attacks between ntp clients and ntp servers. For short AES128 keys, ntpd generates a warning that it is padding them.

Insufficient policy enforcement in navigation in Google Chrome on iOS prior to 90.0.4430.72 allowed a remote attacker to bypass navigation restrictions via a crafted HTML page.

aiohttp is an asynchronous HTTP client/server framework for asyncio and Python. In aiohttp before version 3.7.4 there is an open redirect vulnerability. A maliciously crafted link to an aiohttp-based web-server could redirect the browser to a different website. It is caused by a bug in the `aiohttp.web_middlewares.normalize_path_middleware` middleware. This security problem has been fixed in 3.7.4. Upgrade your dependency using pip as follows "pip install aiohttp >= 3.7.4". If upgrading is not an option for you, a workaround can be to avoid using `aiohttp.web_middlewares.normalize_path_middleware` in your applications.

Flatpak is a system for building, distributing, and running sandboxed desktop applications on Linux. In Flatpack since version 0.9.4 and before version 1.10.2 has a vulnerability in the "file forwarding" feature which can be used by an attacker to gain access to files that would not ordinarily be allowed by the app's permissions. By putting the special tokens `@@` and/or `@@u` in the Exec field of a Flatpak app's .desktop file, a malicious app publisher can trick flatpak into behaving as though the user had chosen to open a target file with their Flatpak app, which automatically makes that file available to the Flatpak app. This is fixed in version 1.10.2. A minimal solution is the first commit "`Disallow @@ and @@U usage in desktop files`". The follow-up commits "`dir: Reserve the whole @@ prefix`" and "`dir: Refuse to export .desktop files with suspicious uses of @@ tokens`" are recommended, but not strictly required. As a workaround, avoid installing Flatpak apps from untrusted sources, or check the contents of the exported `.desktop` files in `exports/share/applications/*.desktop` (typically `~/.local/share/flatpak/exports/share/applications/*.desktop` and `/var/lib/flatpak/exports/share/applications/*.desktop`) to make sure that literal filenames do not follow `@@` or `@@u`.

XStream is a Java library to serialize objects to XML and back again. In XStream before version 1.4.16, there is a vulnerability where the processed stream at unmarshalling time contains type information to recreate the formerly written objects. XStream creates therefore new instances based on these type information. An attacker can manipulate the processed input stream and replace or inject objects, that result in a server-side forgery request. No user is affected, who followed the recommendation to setup XStream's security framework with a whitelist limited to the minimal required types. If you rely on XStream's default blacklist of the Security Framework, you will have to use at least version 1.4.16.

Insufficient policy enforcement in File System API in Google Chrome on Windows prior to 89.0.4389.72 allowed a remote attacker to bypass filesystem restrictions via a crafted HTML page.

Off-by-one error in the get_prng_bytes function in crypto/ansi_cprng.c in the Linux kernel through 3.11.4 makes it easier for context-dependent attackers to defeat cryptographic protection mechanisms via multiple requests for small amounts of data, leading to improper management of the state of the consumed data.

The Python client library for Glance (python-glanceclient) before 0.10.0 does not properly check the preverify_ok value, which prevents the server hostname from being verified with a domain name in the subject's Common Name (CN) or subjectAltName field of the X.509 certificate and allows man-in-the-middle attackers to spoof SSL servers via an arbitrary valid certificate.

A flaw in the processing of received ICMP errors (ICMP fragment needed and ICMP redirect) in the Linux kernel functionality was found to allow the ability to quickly scan open UDP ports. This flaw allows an off-path remote user to effectively bypass the source port UDP randomization. The highest threat from this vulnerability is to confidentiality and possibly integrity, because software that relies on UDP source port randomization are indirectly affected as well.

A flaw was found in mbsync before v1.3.5 and v1.4.1. Validations of the mailbox names returned by IMAP LIST/LSUB do not occur allowing a malicious or compromised server to use specially crafted mailbox names containing '..' path components to access data outside the designated mailbox on the opposite end of the synchronization channel. The highest threat from this vulnerability is to data confidentiality and integrity.

The IcedTea-Web plugin before 1.2.3 and 1.3.x before 1.3.2 uses the same class loader for applets with the same codebase path but from different domains, which allows remote attackers to obtain sensitive information or possibly alter other applets via a crafted applet.

Session fixation vulnerability in Drupal 5.x before 5.9 and 6.x before 6.3, when contributed modules "terminate the current request during a login event," allows remote attackers to hijack web sessions via unknown vectors.

A Security Bypass vulnerability exists in PolarSSL 0.99pre4 through 1.1.1 due to a weak encryption error when generating Diffie-Hellman values and RSA keys.

GnuPG through 2.3.6, in unusual situations where an attacker possesses any secret-key information from a victim's keyring and other constraints (e.g., use of GPGME) are met, allows signature forgery via injection into the status line.

TLS session reuse can lead to host certificate verification bypass in node version < 12.18.0 and < 14.4.0.

Insufficient validation of untrusted input in Omnibox in Google Chrome prior to 80.0.3987.87 allowed a remote attacker to perform domain spoofing via IDN homographs via a crafted domain name.

Insufficient policy enforcement in Blink in Google Chrome prior to 80.0.3987.87 allowed a remote attacker to bypass content security policy via a crafted HTML page.

Insufficient policy enforcement in extensions in Google Chrome prior to 80.0.3987.149 allowed an attacker who convinced a user to install a malicious extension to bypass site isolation via a crafted Chrome Extension.

PyJWT is a Python implementation of RFC 7519. PyJWT supports multiple different JWT signing algorithms. With JWT, an attacker submitting the JWT token can choose the used signing algorithm. The PyJWT library requires that the application chooses what algorithms are supported. The application can specify `jwt.algorithms.get_default_algorithms()` to get support for all algorithms, or specify a single algorithm. The issue is not that big as `algorithms=jwt.algorithms.get_default_algorithms()` has to be used. Users should upgrade to v2.4.0 to receive a patch for this issue. As a workaround, always be explicit with the algorithms that are accepted and expected when decoding.

Vulnerability in the Java SE, Java SE Embedded product of Oracle Java SE (component: Lightweight HTTP Server). Supported versions that are affected are Java SE: 7u251, 8u241, 11.0.6 and 14; Java SE Embedded: 8u241. Difficult to exploit vulnerability allows unauthenticated attacker with network access via multiple protocols to compromise Java SE, Java SE Embedded. Successful attacks of this vulnerability can result in unauthorized update, insert or delete access to some of Java SE, Java SE Embedded accessible data as well as unauthorized read access to a subset of Java SE, Java SE Embedded accessible data. Note: This vulnerability can only be exploited by supplying data to APIs in the specified Component without using Untrusted Java Web Start applications or Untrusted Java applets, such as through a web service. CVSS 3.0 Base Score 4.8 (Confidentiality and Integrity impacts). CVSS Vector: (CVSS:3.0/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N).

In rfb/CSecurityTLS.cxx and rfb/CSecurityTLS.java in TigerVNC before 1.11.0, viewers mishandle TLS certificate exceptions. They store the certificates as authorities, meaning that the owner of a certificate could impersonate any server after a client had added an exception.

Accepting arbitrary Subject Alternative Name (SAN) types, unless a PKI is specifically defined to use a particular SAN type, can result in bypassing name-constrained intermediates. Node.js < 12.22.9, < 14.18.3, < 16.13.2, and < 17.3.1 was accepting URI SAN types, which PKIs are often not defined to use. Additionally, when a protocol allows URI SANs, Node.js did not match the URI correctly.Versions of Node.js with the fix for this disable the URI SAN type when checking a certificate against a hostname. This behavior can be reverted through the --security-revert command-line option.

FreeRDP is a free implementation of the Remote Desktop Protocol (RDP). Prior to version 2.7.0, server side authentication against a `SAM` file might be successful for invalid credentials if the server has configured an invalid `SAM` file path. FreeRDP based clients are not affected. RDP server implementations using FreeRDP to authenticate against a `SAM` file are affected. Version 2.7.0 contains a fix for this issue. As a workaround, use custom authentication via `HashCallback` and/or ensure the `SAM` database path configured is valid and the application has file handles left.

A vulnerability was found in Red Hat Ceph Storage 4 and Red Hat Openshift Container Storage 4.2 where, A nonce reuse vulnerability was discovered in the secure mode of the messenger v2 protocol, which can allow an attacker to forge auth tags and potentially manipulate the data by leveraging the reuse of a nonce in a session. Messages encrypted using a reused nonce value are susceptible to serious confidentiality and integrity attacks.

LibEtPan through 1.9.4, as used in MailCore 2 through 0.6.3 and other products, has a STARTTLS buffering issue that affects IMAP, SMTP, and POP3. When a server sends a "begin TLS" response, the client reads additional data (e.g., from a meddler-in-the-middle attacker) and evaluates it in a TLS context, aka "response injection."

In ectd before versions 3.4.10 and 3.3.23, gateway TLS authentication is only applied to endpoints detected in DNS SRV records. When starting a gateway, TLS authentication will only be attempted on endpoints identified in DNS SRV records for a given domain, which occurs in the discoverEndpoints function. No authentication is performed against endpoints provided in the --endpoints flag. This has been fixed in versions 3.4.10 and 3.3.23 with improved documentation and deprecation of the functionality.

In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, when determining the common dimension size of two tensors, TFLite uses a `DCHECK` which is no-op outside of debug compilation modes. Since the function always returns the dimension of the first tensor, malicious attackers can craft cases where this is larger than that of the second tensor. In turn, this would result in reads/writes outside of bounds since the interpreter will wrongly assume that there is enough data in both tensors. The issue is patched in commit 8ee24e7949a203d234489f9da2c5bf45a7d5157d, and is released in TensorFlow versions 1.15.4, 2.0.3, 2.1.2, 2.2.1, or 2.3.1.

In TensorFlow Lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, saved models in the flatbuffer format use a double indexing scheme: a model has a set of subgraphs, each subgraph has a set of operators and each operator has a set of input/output tensors. The flatbuffer format uses indices for the tensors, indexing into an array of tensors that is owned by the subgraph. This results in a pattern of double array indexing when trying to get the data of each tensor. However, some operators can have some tensors be optional. To handle this scenario, the flatbuffer model uses a negative `-1` value as index for these tensors. This results in special casing during validation at model loading time. Unfortunately, this means that the `-1` index is a valid tensor index for any operator, including those that don't expect optional inputs and including for output tensors. Thus, this allows writing and reading from outside the bounds of heap allocated arrays, although only at a specific offset from the start of these arrays. This results in both read and write gadgets, albeit very limited in scope. The issue is patched in several commits (46d5b0852, 00302787b7, e11f5558, cd31fd0ce, 1970c21, and fff2c83), and is released in TensorFlow versions 1.15.4, 2.0.3, 2.1.2, 2.2.1, or 2.3.1. A potential workaround would be to add a custom `Verifier` to the model loading code to ensure that only operators which accept optional inputs use the `-1` special value and only for the tensors that they expect to be optional. Since this allow-list type approach is erro-prone, we advise upgrading to the patched code.

By exploiting an Open Redirect vulnerability on a website, an attacker could have spoofed the site displayed in the download file dialog to show the original site (the one suffering from the open redirect) rather than the site the file was actually downloaded from. This vulnerability affects Firefox < 81, Thunderbird < 78.3, and Firefox ESR < 78.3.

GnuTLS 3.6.x before 3.6.14 uses incorrect cryptography for encrypting a session ticket (a loss of confidentiality in TLS 1.2, and an authentication bypass in TLS 1.3). The earliest affected version is 3.6.4 (2018-09-24) because of an error in a 2018-09-18 commit. Until the first key rotation, the TLS server always uses wrong data in place of an encryption key derived from an application.

Vulnerability in the Java SE, Java SE Embedded product of Oracle Java SE (component: Libraries). Supported versions that are affected are Java SE: 8u251, 11.0.7 and 14.0.1; Java SE Embedded: 8u251. Difficult to exploit vulnerability allows unauthenticated attacker with network access via multiple protocols to compromise Java SE, Java SE Embedded. Successful attacks of this vulnerability can result in unauthorized update, insert or delete access to some of Java SE, Java SE Embedded accessible data as well as unauthorized read access to a subset of Java SE, Java SE Embedded accessible data. Note: Applies to client and server deployment of Java. This vulnerability can be exploited through sandboxed Java Web Start applications and sandboxed Java applets. It can also be exploited by supplying data to APIs in the specified Component without using sandboxed Java Web Start applications or sandboxed Java applets, such as through a web service. CVSS 3.1 Base Score 4.8 (Confidentiality and Integrity impacts). CVSS Vector: (CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N).