JinJava sandbox bypass via ForTag leading to arbitrary Java execution
In JinJava (HubSpot) versions prior to 2.7.6 and 2.8.3, a sandbox bypass in template rendering allows arbitrary Java execution. The bypass is reachable via the ForTag implementation used for iterating over lists/items in templates: ForTag iterates over object properties using Java bean introspection (Introspector.getBeanInfo()) without enforcing JinJava’s Bean ELResolver restrictions, enabling invocation of getter methods that should be blocked by the sandbox. In reported exploitation chains, this can be combined with an ObjectMapper JSON deserialization path to bypass the sandbox type allowlist and instantiate otherwise-restricted classes (e.g., JinjavaConfig or JinjavaELContext), effectively weakening or rewriting sandbox rules. The net result is arbitrary Java class instantiation and file access despite built-in sandbox restrictions.
What it means and what to do now: patch path, interim mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
Successful exploitation allows an attacker to bypass JinJava’s safety sandbox and achieve arbitrary Java execution on the server. This includes arbitrary Java class instantiation and server-side file access (potentially reading sensitive files), and can lead to broader compromise outcomes such as data modification or service disruption/crash, with high impact to confidentiality, integrity, and availability.
If you can’t patch tonight, do this now.
If immediate upgrade is not possible, do not render untrusted/user-provided templates with JinJava. Restrict and audit who can create/edit templates, minimize/remove powerful Java objects from the template context, and prevent template authors from accessing ObjectMapper/default typing or other high-risk objects. Treat any user-controlled template rendering environment as high risk and isolate/segment it to limit blast radius.
Patch, then assume compromise.
Upgrade JinJava to a patched version: 2.7.6 or later, or 2.8.3 or later. The fixes include additional security checks in ForTag rendering and tightened ObjectMapper behavior / restricted-type validation to prevent type allowlist bypass and sandbox weakening.
No validated public exploits yet. Hidden: 1 rejected by Mallory's analysis as not a working exploit.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
10 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.