Description
Arbitrary Class Instantiation via XML Feature Generator Descriptor and Format Name in Apache OpenNLP

Versions Affected:

- before 2.5.10
- before 3.0.0-M5

Description:

Three code paths in Apache OpenNLP load a class by its fully-qualified name via Class.forName() and invoke its no-arg constructor without any prior validation of the class name or its type. 

The affected paths are:

(1) GeneratorFactory, which reads the class attribute of generator elements in an XML feature generator descriptor; such descriptors are embedded as artifacts in model archives (e.g. TokenNameFinder and POSTagger models) and are parsed during model loading, so an attacker who can supply a crafted model archive controls the class name directly.

(2) StreamFactoryRegistry.getFactory(Class, String), which falls back to interpreting an unregistered format name as the fully-qualified class name of an ObjectStreamFactory; this is exploitable in applications that pass untrusted format names (e.g. exposing the -format parameter of the command-line tooling to external input).

(3) StringInterners, which instantiates the interner implementation named by the opennlp.interner.class system property; this value is normally deployer-controlled, so it is hardened as defense in depth rather than being independently attacker-reachable.

Exploitation requires a class with attacker-useful side effects in its static initializer or no-arg constructor (JNDI lookup, outbound network I/O, filesystem access) to be present on the classpath, so this is not drop-in remote code execution. T

Mitigation:

Upgrade to a fixed release.

The fix routes all three paths through ExtensionLoader.instantiateExtension(...), which consults a package-prefix allowlist before Class.forName() is invoked, so a disallowed class is never loaded, initialized, or constructed.
Classes under the opennlp. prefix remain permitted by default. Deployments that load models referencing feature generator factories, object stream factories, or string interners outside opennlp.* must opt those packages in, either programmatically via ExtensionLoader.registerAllowedPackage(String) before the first model load, or by setting the OPENNLP_EXT_ALLOWED_PACKAGES system property to a comma-separated list of allowed package prefixes.

Users who cannot upgrade immediately should ensure all model files and format names are sourced from trusted origins and should audit their classpath for classes with side-effecting static initializers or constructors.
Published: 2026-07-24
Score: 5.6 Medium
EPSS: < 1% Very Low
KEV: No
Impact: n/a
Action: n/a
AI Analysis

Impact

Apache OpenNLP allows a user-supplied XML descriptor or untrusted format name to specify any fully‑qualified class that the Java runtime will load and instantiate through a no‑argument constructor. The loaded class is not validated against a whitelist before construction, enabling an attacker to execute arbitrary code during model loading or object stream creation. The vulnerability does not directly provide remote code execution; it requires that a malicious class with side‑effecting static initializers or constructors be present on the application’s classpath, but if such a class is available it can perform JNDI lookups, outbound network traffic, or file‑system operations, compromising confidentiality, integrity, or availability.

Affected Systems

The flaw exists in all Apache OpenNLP releases prior to 2.5.10 and all releases before 3.0.0-M5. The product is released by the Apache Software Foundation under the Apache OpenNLP project.

Risk and Exploitability

The CVSS score of 5.6 indicates a moderate impact level. The EPSS score of less than 1% implies a very low probability of exploitation in the wild, and the vulnerability is not listed in the CISA KEV catalog. The attack vector is inferred to be through the ingestion of a malicious model archive or an externally supplied format name; a successful exploit requires the attacker to supply a crafted model file or control the value of the command-line -format option so that the runtime can construct a disallowed class from the classpath.

Generated by OpenCVE AI on August 3, 2026 at 20:31 UTC.

Remediation

No vendor fix or workaround currently provided.

OpenCVE Recommended Actions

  • Upgrade Apache OpenNLP to version 2.5.10 or later, or 3.0.0-M5 or later, where the unsafe instantiation paths are replaced by a package‑prefix allowlist.
  • If a version upgrade cannot be performed immediately, ensure that all model archives and format names are sourced from trusted, signed origins and that no untrusted models are loaded by the application.
  • Audit the application classpath for classes with side‑effecting static initializers or constructors and remove or neutralize any that could be exploited.
  • When using a fixed release that permits optional packages, call ExtensionLoader.registerAllowedPackage("package.name") before model loading, or set the OPENNLP_EXT_ALLOWED_PACKAGES system property to a comma‑separated list of allowed package prefixes to constrain the set of classes that may be instantiated by future feature descriptors or stream factories.

Generated by OpenCVE AI on August 3, 2026 at 20:31 UTC.

Tracking

Sign in to view the affected projects.

Advisories

No advisories yet.

History

Fri, 24 Jul 2026 19:30:00 +0000

Type Values Removed Values Added
Metrics cvssV3_1

{'score': 5.6, 'vector': 'CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:L'}

ssvc

{'options': {'Automatable': 'no', 'Exploitation': 'none', 'Technical Impact': 'partial'}, 'version': '2.0.3'}

Fri, 24 Jul 2026 18:30:00 +0000

Type Values Removed Values Added
First Time appeared Apache
Apache opennlp
Vendors & Products Apache
Apache opennlp

Fri, 24 Jul 2026 08:15:00 +0000

Type Values Removed Values Added
Description Arbitrary Class Instantiation via XML Feature Generator Descriptor and Format Name in Apache OpenNLP Versions Affected: - before 2.5.10 - before 3.0.0-M5 Description: Three code paths in Apache OpenNLP load a class by its fully-qualified name via Class.forName() and invoke its no-arg constructor without any prior validation of the class name or its type.  The affected paths are: (1) GeneratorFactory, which reads the class attribute of generator elements in an XML feature generator descriptor; such descriptors are embedded as artifacts in model archives (e.g. TokenNameFinder and POSTagger models) and are parsed during model loading, so an attacker who can supply a crafted model archive controls the class name directly. (2) StreamFactoryRegistry.getFactory(Class, String), which falls back to interpreting an unregistered format name as the fully-qualified class name of an ObjectStreamFactory; this is exploitable in applications that pass untrusted format names (e.g. exposing the -format parameter of the command-line tooling to external input). (3) StringInterners, which instantiates the interner implementation named by the opennlp.interner.class system property; this value is normally deployer-controlled, so it is hardened as defense in depth rather than being independently attacker-reachable. Exploitation requires a class with attacker-useful side effects in its static initializer or no-arg constructor (JNDI lookup, outbound network I/O, filesystem access) to be present on the classpath, so this is not drop-in remote code execution. T Mitigation: Upgrade to a fixed release. The fix routes all three paths through ExtensionLoader.instantiateExtension(...), which consults a package-prefix allowlist before Class.forName() is invoked, so a disallowed class is never loaded, initialized, or constructed. Classes under the opennlp. prefix remain permitted by default. Deployments that load models referencing feature generator factories, object stream factories, or string interners outside opennlp.* must opt those packages in, either programmatically via ExtensionLoader.registerAllowedPackage(String) before the first model load, or by setting the OPENNLP_EXT_ALLOWED_PACKAGES system property to a comma-separated list of allowed package prefixes. Users who cannot upgrade immediately should ensure all model files and format names are sourced from trusted origins and should audit their classpath for classes with side-effecting static initializers or constructors.
Title Apache OpenNLP: Arbitrary Class Instantiation in GeneratorFactory via Feature Descriptor XML
Weaknesses CWE-470
References

Subscriptions

Apache Opennlp
cve-icon MITRE

Status: PUBLISHED

Assigner: apache

Published:

Updated: 2026-07-24T18:04:05.691Z

Reserved: 2026-07-16T13:23:13.170Z

Link: CVE-2026-63317

cve-icon Vulnrichment

Updated: 2026-07-24T14:34:49.489Z

cve-icon NVD

Status : Analyzed

Published: 2026-07-24T09:16:25.273

Modified: 2026-08-06T00:48:32.980

Link: CVE-2026-63317

cve-icon Redhat

No data.

cve-icon OpenCVE Enrichment

Updated: 2026-08-03T20:45:03Z

Weaknesses
  • CWE-470

    Use of Externally-Controlled Input to Select Classes or Code ('Unsafe Reflection')