GHSA-M49C-G9WR-HV6V
Vulnerability from github – Published: 2025-09-17 19:56 – Updated: 2026-07-31 22:15Summary
jinjava’s current sandbox restrictions prevent direct access to dangerous methods such as getClass(), and block instantiation of Class objects. However, these protections can be bypassed.
By using mapper.getTypeFactory().constructFromCanonical(), it is possible to instruct the underlying ObjectMapper to deserialize attacker-controlled input into arbitrary classes. This enables the creation of semi-arbitrary class instances without directly invoking restricted methods or class literals.
As a result, an attacker can escape the sandbox and instantiate classes such as java.net.URL, opening up the ability to access local files and URLs(e.g., file:///etc/passwd). With further chaining, this primitive can potentially lead to remote code execution (RCE).
Details
jinjava templates expose a built-in variable ____int3rpr3t3r____, which provides direct access to the jinjavaInterpreter instance.
This variable was previously abused and protections were added to prevent call method from JinjavaInterpreter instances (see Add interpreter to blacklist).
However, interacting with the properties of JinjavaInterpreter instances remains unrestricted.
From ____int3rpr3t3r____, it is possible to traverse to the config field, which exposes an ObjectMapper. By invoking readValue(String content, JavaType valueType) on this ObjectMapper, an attacker can instantiate arbitrary classes specified via JavaType.
Although jinjava explicitly restricts dangerous classes such as Class, ClassLoader, and so on inside JinjavaBeanELResolver, the JavaType class itself is not restricted.
As a result, an attacker can leverage JavaType construction (constructFromCanonical) to instantiate semi-arbitrary classes without directly calling restricted methods.
This allows sandbox escape and the creation of powerful primitives.
Impact
Escape the Jinjava sandbox and instantiate a wide range of classes using JavaType. This capability can be used to read arbitrary files and to perform full read SSRF by creating network-related objects. In certain environments, depending on the available classes, this primitive can even lead to complete remote code execution.
{
"affected": [
{
"package": {
"ecosystem": "Maven",
"name": "com.hubspot.jinjava:jinjava"
},
"ranges": [
{
"events": [
{
"introduced": "2.8.0"
},
{
"fixed": "2.8.1"
}
],
"type": "ECOSYSTEM"
}
],
"versions": [
"2.8.0"
]
},
{
"package": {
"ecosystem": "Maven",
"name": "com.hubspot.jinjava:jinjava"
},
"ranges": [
{
"events": [
{
"introduced": "2.7.0"
},
{
"fixed": "2.7.5"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2025-59340"
],
"database_specific": {
"cwe_ids": [
"CWE-1336"
],
"github_reviewed": true,
"github_reviewed_at": "2025-09-17T19:56:21Z",
"nvd_published_at": "2025-09-17T20:15:36Z",
"severity": "CRITICAL"
},
"details": "### Summary\n\njinjava\u2019s current sandbox restrictions prevent direct access to dangerous methods such as `getClass()`, and block instantiation of Class objects. However, these protections can be bypassed.\n\nBy using mapper.getTypeFactory().constructFromCanonical(), it is possible to instruct the underlying ObjectMapper to deserialize attacker-controlled input into arbitrary classes. This enables the creation of semi-arbitrary class instances without directly invoking restricted methods or class literals.\n\nAs a result, an attacker can escape the sandbox and instantiate classes such as java.net.URL, opening up the ability to access local files and URLs(e.g., file:///etc/passwd). With further chaining, this primitive can potentially lead to remote code execution (RCE).\n\n### Details\n\njinjava templates expose a built-in variable `____int3rpr3t3r____`, which provides direct access to the jinjavaInterpreter instance.\nThis variable was previously abused and protections were added to prevent call method from `JinjavaInterpreter` instances (see [Add interpreter to blacklist](https://github.com/HubSpot/jinjava/commit/1b9aaa4b420c58b4a301cf4b7d26207f1c8d1165)).\nHowever, interacting with the properties of `JinjavaInterpreter` instances remains [unrestricted](https://github.com/HubSpot/jinjava/blob/jinjava-2.8.0/src/main/java/com/hubspot/jinjava/el/ext/JinjavaBeanELResolver.java#L80-L84).\n\nFrom `____int3rpr3t3r____`, it is possible to traverse to the `config` field, which exposes an ObjectMapper. By invoking `readValue(String content, JavaType valueType)` on this ObjectMapper, an attacker can instantiate arbitrary classes specified via `JavaType`.\n\nAlthough jinjava explicitly restricts dangerous classes such as `Class`, `ClassLoader`, and so on inside `JinjavaBeanELResolver`, the `JavaType` class itself is [not restricted](https://github.com/HubSpot/jinjava/blob/jinjava-2.8.0/src/main/java/com/hubspot/jinjava/el/ext/JinjavaBeanELResolver.java#L246-L262). \n\nAs a result, an attacker can leverage `JavaType` construction (`constructFromCanonical`) to instantiate semi-arbitrary classes without directly calling restricted methods.\n\nThis allows sandbox escape and the creation of powerful primitives. \n\n### Impact\nEscape the Jinjava sandbox and instantiate a wide range of classes using JavaType.\nThis capability can be used to read arbitrary files and to perform full read SSRF by creating network-related objects.\nIn certain environments, depending on the available classes, this primitive can even lead to complete remote code execution.",
"id": "GHSA-m49c-g9wr-hv6v",
"modified": "2026-07-31T22:15:57Z",
"published": "2025-09-17T19:56:21Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/HubSpot/jinjava/security/advisories/GHSA-m49c-g9wr-hv6v"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2025-59340"
},
{
"type": "WEB",
"url": "https://github.com/github/advisory-database/pull/8141"
},
{
"type": "WEB",
"url": "https://github.com/HubSpot/jinjava/commit/66df351e7e8ad71ca04dcacb4b65782af820b8b1"
},
{
"type": "PACKAGE",
"url": "https://github.com/HubSpot/jinjava"
},
{
"type": "WEB",
"url": "https://github.com/HubSpot/jinjava/releases/tag/jinjava-2.7.5"
},
{
"type": "WEB",
"url": "https://github.com/HubSpot/jinjava/releases/tag/jinjava-2.8.1"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
],
"summary": "jinjava has Sandbox Bypass via JavaType-Based Deserialization"
}
Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
Vulnerability timeline
Sightings and recorded dates, grouped by day · UTC
How durations are calculated
Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.
Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.
Loading timeline…
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.