CVE-2026-42027
CriticalCVSS 9.8Exploitation Probability (EPSS)
Low risk50th percentile - higher than 50% of all known CVEs
Summary
In Apache OpenNLP before versions 1.9.5, 2.5.9, and 3.0.0-M3, the ExtensionLoader.instantiateExtension() method loads a class by name from a model archive's manifest.properties using Class.forName() before checking type compatibility. This allows an attacker to execute the static initializer of any class on the classpath during model loading, potentially causing harmful side effects.
Risk Assessment
The risk is potential remote code execution (RCE) or triggering malicious actions like JNDI lookups, network traffic, or file operations if an attacker supplies a crafted model. This is especially dangerous in environments where models are loaded from untrusted sources, such as community repositories.
Recommendation
Upgrade Apache OpenNLP to version 2.5.9 (for 2.x) or 3.0.0-M3 (for 3.x) immediately. If upgrading is not possible, only load models from trusted sources and audit the classpath for classes with dangerous static initializers.
Other vulnerabilities in Apache OpenNLP
See all- CVE-2026-82617Critical
The built-in name-finder patterns EMAIL and URL in Apache OpenNLP contain ambiguous nested quantifiers, enabling a ReDoS attack. A crafted input can trigger super-linear backtracking or unbounded matcher recursion, leading to CPU exhaustion or StackOverflowError. The issue affects versions 2.0.0 through 2.5.11 and 3.0.0-M1 through 3.0.0-M5.
- CVE-2026-43825High
In Apache OpenNLP, the libsvm document categorization module (3.x line) is vulnerable to untrusted Java deserialization. The SvmDoccatModel.deserialize(InputStream) method uses java.io.ObjectInputStream without a filter, allowing an attacker to execute arbitrary code via a crafted stream.
- CVE-2026-40682Critical
An XML External Entity (XXE) vulnerability was found in Apache OpenNLP's DictionaryEntryPersistor class. The class initializes a SAX parser without enabling secure processing or disabling DTD handling, allowing an attacker to inject a malicious DOCTYPE declaration in a dictionary file. This can lead to local file disclosure via file:// entity references or server-side request forgery via http:// entity references during XML parsing.
- CVE-2026-67211High
A vulnerability in Apache OpenNLP (opennlp-spellcheck extension, versions 3.0.0-M4 and 3.0.0-M5) allows an Out-of-Memory Denial of Service. The SymSpellModelSerializer.create() method reads unigram and bigram count fields from a model file and passes them directly to map pre-sizing without an upper bound, causing a crafted .bin file to attempt a 4–8 GB allocation and fail with OutOfMemoryError.
- CVE-2026-63317Medium
In Apache OpenNLP before versions 2.5.10 and 3.0.0-M5, a vulnerability allows arbitrary class instantiation. An attacker can supply a crafted model or format name, leading to loading and executing code from dangerous classes.
- CVE-2026-42440High
In Apache OpenNLP before versions 1.9.5, 2.5.9, and 3.0.0-M3, a denial-of-service vulnerability exists due to unbounded array allocation in AbstractModelReader. The methods getOutcomes(), getOutcomePatterns(), and getPredicates() allocate arrays based on a 32-bit integer from the model file without validation, allowing an attacker to set the value to Integer.MAX_VALUE and trigger an OutOfMemoryError.
Original NVD description (English source)
Arbitrary Class Instantiation via Model Manifest in Apache OpenNLP ExtensionLoader Versions Affected: before 1.9.5, before 2.5.9, before 3.0.0-M3 Description: The ExtensionLoader.instantiateExtension(Class, String) method loads a class by its fully-qualified name via Class.forName() and invokes its no-arg constructor, with the class name sourced from the manifest.properties entry of a model archive. The existing isAssignableFrom check correctly rejects classes that are not subtypes of the expected extension interface (BaseToolFactory for factory=, ArtifactSerializer for serializer-class-*), but the check runs after Class.forName() has already loaded and initialized the named class. Class.forName() with default initialization semantics executes the target class's static initializer before returning, so an attacker who can supply a crafted model archive can cause the static initializer of any class on the classpath to run during model loading, regardless of whether that class passes the subsequent type check. Exploitation requires a class with attacker-useful side effects in its static initializer (for example, JNDI lookup, outbound network I/O, or filesystem access) to be present on the classpath, so this is not a drop-in remote code execution; however, the attack surface grows as third-party model distribution becomes more common (community model repositories, Hugging Face-style sharing), where users routinely load model files from origins they do not control. A secondary, narrower vector affects deployments that ship legitimate BaseToolFactory or ArtifactSerializer subclasses with side-effecting no-arg constructors: a malicious manifest can name such a class and force its constructor to run during model load. Mitigation: * 2.x users should upgrade to 2.5.9. * 3.x users should upgrade to 3.0.0-M3. Note: The fix introduces a package-prefix allowlist that is consulted before Class.forName() is invoked, so the static initializer of a disallowed class is never executed. Classes under the opennlp. prefix remain permitted by default. Deployments that load models referencing factories or serializers 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 that all model files are sourced from trusted origins and should audit their classpath for classes with side-effecting static initializers or constructors, particularly any that perform JNDI lookups, network requests, or filesystem operations during class initialization.
Vulnerability data from NVD (NIST) · CISA KEV · EPSS