http.cookiejar.DefaultPolicy.domain_return_ok in Lib/http/cookiejar.py in Python before 3.7.3 does not correctly validate the domain: it can be tricked into sending existing cookies to the wrong server. An attacker may abuse this flaw by using a server with a hostname that has another valid hostname as a suffix (e.g., pythonicexample.com to steal cookies for example.com). When a program uses http.cookiejar.DefaultPolicy and tries to do an HTTP connection to an attacker-controlled server, existing cookies can be leaked to the attacker. This affects 2.x through 2.7.16, 3.x before 3.4.10, 3.5.x before 3.5.7, 3.6.x before 3.6.9, and 3.7.x before 3.7.3.

Rapid7 Insight Agent, version 2.6.3 and prior, suffers from a local privilege escalation due to an uncontrolled DLL search path. Specifically, when Insight Agent 2.6.3 and prior starts, the Python interpreter attempts to load python3.dll at "C:\DLLs\python3.dll," which normally is writable by locally authenticated users. Because of this, a malicious local user could use Insight Agent's startup conditions to elevate to SYSTEM privileges. This issue was fixed in Rapid7 Insight Agent 2.6.4.

The MSI installer for Python through 2.7.16 on Windows defaults to the C:\Python27 directory, which makes it easier for local users to deploy Trojan horse code. (This also affects old 3.x releases before 3.5.) NOTE: the vendor's position is that it is the user's responsibility to ensure C:\Python27 access control or choose a different directory, because backwards compatibility requires that C:\Python27 remain the default for 2.7.x

An issue was discovered in Riello NetMan 204 14-2 and 15-2. The issue is with the login script and wrongpass Python script used for authentication. When calling wrongpass, the variables $VAL0 and $VAL1 should be enclosed in quotes to prevent the potential for Bash command injection. Further to this, VAL0 and VAL1 should be sanitised to ensure they do not contain malicious characters. Passing it the username of '-' will cause it to time out and log the user in because of poor error handling. This will log the attacker in as an administrator where the telnet / ssh services can be enabled, and the credentials for local users can be reset. Also, login.cgi accepts the username as a GET parameter, so login can be achieved by browsing to the /cgi-bin/login.cgi?username=-%20a URI.

The HTTP API supported by Starry Station (aka Starry Router) allows brute forcing the PIN setup by the user on the device, and this allows an attacker to change the Wi-Fi settings and PIN, as well as port forward and expose any internal device's port to the Internet. It was identified that the device uses custom Python code called "rodman" that allows the mobile appication to interact with the device. The APIs that are a part of this rodman Python file allow the mobile application to interact with the device using a secret, which is a uuid4 based session identifier generated by the device the first time it is set up. However, in some cases, these APIs can also use a security code. This security code is nothing but the PIN number set by the user to interact with the device when using the touch interface on the router. This allows an attacker on the Internet to interact with the router's HTTP interface when a user navigates to the attacker's website, and brute force the credentials. Also, since the device's server sets the Access-Control-Allow-Origin header to "*", an attacker can easily interact with the JSON payload returned by the device and steal sensitive information about the device.

A security regression of CVE-2019-9636 was discovered in python since commit d537ab0ff9767ef024f26246899728f0116b1ec3 affecting versions 2.7, 3.5, 3.6, 3.7 and from v3.8.0a4 through v3.8.0b1, which still allows an attacker to exploit CVE-2019-9636 by abusing the user and password parts of a URL. When an application parses user-supplied URLs to store cookies, authentication credentials, or other kind of information, it is possible for an attacker to provide specially crafted URLs to make the application locate host-related information (e.g. cookies, authentication data) and send them to a different host than where it should, unlike if the URLs had been correctly parsed. The result of an attack may vary based on the application.

A code injection issue was discovered in PyXDG before 0.26 via crafted Python code in a Category element of a Menu XML document in a .menu file. XDG_CONFIG_DIRS must be set up to trigger xdg.Menu.parse parsing within the directory containing this file. This is due to a lack of sanitization in xdg/Menu.py before an eval call.

Prima Systems FlexAir, Versions 2.4.9api3 and prior. The application allows the upload of arbitrary Python scripts when configuring the main central controller. These scripts can be immediately executed because of root code execution, not as a web server user, allowing an authenticated attacker to gain full system access.

A vulnerability in the Python scripting subsystem of Cisco NX-OS Software could allow an authenticated, local attacker to escape the Python parser and issue arbitrary commands to elevate the attacker's privilege level. The vulnerability is due to insufficient sanitization of user-supplied parameters that are passed to certain Python functions in the scripting sandbox of the affected device. An attacker could exploit this vulnerability to escape the scripting sandbox and execute arbitrary commands to elevate the attacker's privilege level. To exploit this vulnerability, the attacker must have local access and be authenticated to the targeted device with administrative or Python execution privileges. These requirements could limit the possibility of a successful exploit.

Python Facebook Thrift servers would not error upon receiving messages with containers of fields of unknown type. As a result, malicious clients could send short messages which would take a long time for the server to parse, potentially leading to denial of service. This issue affects Facebook Thrift prior to v2019.02.18.00.

Any Python module in sys.path can be imported if the command line of the process triggering the coredump is Python and the first argument is -m in Apport before 2.19.2 function _python_module_path.

python-dbusmock before version 0.15.1 AddTemplate() D-Bus method call or DBusTestCase.spawn_server_template() method could be tricked into executing malicious code if an attacker supplies a .pyc file.

util/emailutils.py in Matrix Sydent before 1.0.2 mishandles registration restrictions that are based on e-mail domain, if the allowed_local_3pids option is enabled. This occurs because of potentially unwanted behavior in Python, in which an email.utils.parseaddr call on user@bad.example.net@good.example.com returns the user@bad.example.net substring.

The urllib3 library before 1.24.2 for Python mishandles certain cases where the desired set of CA certificates is different from the OS store of CA certificates, which results in SSL connections succeeding in situations where a verification failure is the correct outcome. This is related to use of the ssl_context, ca_certs, or ca_certs_dir argument.

OneLogin PythonSAML 2.3.0 and earlier may incorrectly utilize the results of XML DOM traversal and canonicalization APIs in such a way that an attacker may be able to manipulate the SAML data without invalidating the cryptographic signature, allowing the attack to potentially bypass authentication to SAML service providers.

In the urllib3 library through 1.24.1 for Python, CRLF injection is possible if the attacker controls the request parameter.

An eval injection vulnerability in the Python web server routing on the Zyxel NAS 326 version 5.21 and below allows a remote authenticated attacker to execute arbitrary code via the tjp6jp6y4, simZysh, and ck6fup6 APIs.

It was found that libreoffice before versions 6.0.7 and 6.1.3 was vulnerable to a directory traversal attack which could be used to execute arbitrary macros bundled with a document. An attacker could craft a document, which when opened by LibreOffice, would execute a Python method from a script in any arbitrary file system location, specified relative to the LibreOffice install location.

urllib in Python 2.x through 2.7.16 supports the local_file: scheme, which makes it easier for remote attackers to bypass protection mechanisms that blacklist file: URIs, as demonstrated by triggering a urllib.urlopen('local_file:///etc/passwd') call.

An issue was discovered in urllib2 in Python 2.x through 2.7.16 and urllib in Python 3.x through 3.7.3. CRLF injection is possible if the attacker controls a url parameter, as demonstrated by the first argument to urllib.request.urlopen with \r\n (specifically in the path component of a URL that lacks a ? character) followed by an HTTP header or a Redis command. This is similar to the CVE-2019-9740 query string issue. This is fixed in: v2.7.17, v2.7.17rc1, v2.7.18, v2.7.18rc1; v3.5.10, v3.5.10rc1, v3.5.8, v3.5.8rc1, v3.5.8rc2, v3.5.9; v3.6.10, v3.6.10rc1, v3.6.11, v3.6.11rc1, v3.6.12, v3.6.9, v3.6.9rc1; v3.7.4, v3.7.4rc1, v3.7.4rc2, v3.7.5, v3.7.5rc1, v3.7.6, v3.7.6rc1, v3.7.7, v3.7.7rc1, v3.7.8, v3.7.8rc1, v3.7.9.