Threats affecting pkg:pypi/jinja2

Jinja2 versions prior to 3.1.6 contain a sandbox breakout vulnerability via the attr filter, allowing attackers controlling template content to execute arbitrary Python code. The flaw arises because the attr filter can bypass sandbox restrictions on the string format method. This vulnerability affects applications that execute untrusted templates. A fix is available in version 3.1.6 that prevents the attr filter from bypassing sandbox attribute lookups.

CVE-2024-56326 is a high-severity vulnerability in Jinja2 versions prior to 3.1.5 that allows sandbox breakout via indirect calls to the str.format method. An attacker who can control template content in applications executing untrusted templates may exploit this to execute arbitrary Python code. The vulnerability arises because the sandbox does not detect indirect references to str.format, which can be invoked through custom filters. This issue is fixed in Jinja2 version 3.1.5.

Red Hat OpenShift Container Platform 4.15.25 includes a security update addressing a vulnerability in the jinja2 package (CVE-2024-34064) where keys containing non-attribute characters are accepted. This issue has been rated as moderate in severity by Red Hat Product Security. Users of OpenShift Container Platform 4.15 are advised to upgrade to the updated packages and container images to mitigate this vulnerability.

Jinja2 versions prior to 3.1.3 have a vulnerability in the xmlattr filter that allows HTML attribute injection when user input is passed as keys. This occurs because keys containing spaces are accepted, which can be interpreted as multiple attributes, enabling potential cross-site scripting (XSS) if untrusted keys are rendered. This issue is mitigated by verifying keys or upgrading to version 3.1.3 or later.