Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB
CVE-2026-111868 Overview
CVE-2026-111868 is an insecure deserialization vulnerability in DataStax Pulsar JMS versions up to 9.0.3. The flaw resides in the decode function of pulsar-jms/src/main/java/com/datastax/oss/pulsar/jms/messages/PulsarObjectMessage.java. An authenticated remote attacker can submit crafted serialized Java objects that the client deserializes without sufficient validation. The weakness is tracked as improper input validation [CWE-20]. Public disclosure of the exploitation technique has occurred, and the project has not responded to the initial issue report. The vulnerability affects confidentiality, integrity, and availability at a limited scope.
Critical Impact
Remote attackers with low privileges can trigger unsafe Java object deserialization through PulsarObjectMessage.decode, enabling tampering with message processing in Pulsar JMS clients.
Affected Products
- DataStax Pulsar JMS versions up to and including 9.0.3
- Applications embedding the pulsar-jms client library
- Downstream Apache Pulsar deployments consuming object messages via this library
Discovery Timeline
- 2026-10-18 - CVE CVE-2026-111868 published to NVD
- 2026-10-19 - Last updated in NVD database
Technical Details for CVE-2026-111868
Vulnerability Analysis
The vulnerability exists in the decode method of PulsarObjectMessage, the class responsible for converting the binary payload of a JMS ObjectMessage back into a Java object. The implementation performs Java native deserialization on attacker-controlled bytes without restricting the set of classes allowed to be instantiated.
When a Pulsar consumer processes a message produced by an attacker who can publish to a topic, the receiving JVM constructs objects defined by the serialized stream. Gadget chains present in the application classpath can be chained to influence program state during object reconstruction. The vulnerability affects any consumer configured to receive object messages from topics an attacker can reach.
The public disclosure increases practical risk. The DataStax project has not yet acknowledged the issue report referenced in pulsar-jms/issues/210, so no vendor-coordinated fix exists at the time of writing.
Root Cause
The root cause is the use of unrestricted Java deserialization in PulsarObjectMessage.decode. The method accepts the raw payload and invokes standard ObjectInputStream deserialization without a class allow-list, filter, or signature check. This matches the pattern described by [CWE-20] improper input validation and the broader class of insecure Java deserialization flaws.
Attack Vector
The attack path is network-based and requires low privileges. An attacker who can produce messages to a Pulsar topic consumed by a vulnerable pulsar-jms client submits a crafted serialized object. When the consumer calls getObject() on the received message, decode deserializes the payload, triggering gadget execution during object graph reconstruction. No user interaction is required on the consumer side.
No verified proof-of-concept code is published in the referenced advisories. Readers should consult the GitHub Issue #210 on Pulsar JMS and VulDB CVE-2026-111868 Details for technical context.
Stop Remote Code Execution Before It Runs
Singularity™ Endpoint detects and blocks exploit attempts at machine speed, including zero-days and novel attack techniques.
Stop Exploit AttemptsDetection Methods for CVE-2026-111868
Indicators of Compromise
- Unexpected class loading events in Pulsar JMS consumer JVMs, especially classes outside normal application packages
- Pulsar ObjectMessage payloads containing Java serialization magic bytes (AC ED 00 05) from untrusted producers
- Consumer process crashes or stack traces referencing ObjectInputStream.readObject within PulsarObjectMessage.decode
Detection Strategies
- Instrument Pulsar JMS consumers with a JVM deserialization filter (jdk.serialFilter) that logs or rejects unexpected classes
- Inspect Pulsar topic traffic for object messages originating from producers not on an approved allow-list
- Correlate consumer-side anomalies with producer identities using Pulsar authentication logs
Monitoring Recommendations
- Enable JVM-level audit logging for deserialization and alert on classes outside the application allow-list
- Monitor Pulsar broker authorization logs for new or unexpected producers on topics carrying object messages
- Track dependency inventories for the pulsar-jms artifact at versions ≤ 9.0.3 across build pipelines
How to Mitigate CVE-2026-111868
Immediate Actions Required
- Identify all services bundling pulsar-jms ≤ 9.0.3 and inventory topics using ObjectMessage
- Restrict producer permissions on affected topics to authenticated, trusted principals only
- Replace ObjectMessage usage with BytesMessage or TextMessage carrying JSON or Protobuf where feasible
Patch Information
No vendor patch is available at the time of publication. The project maintainers have not responded to the referenced issue report. Monitor the GitHub Pulsar JMS Repository and GitHub Issue #210 on Pulsar JMS for remediation updates.
Workarounds
- Apply a JVM-wide serialization filter via -Djdk.serialFilter restricting deserialization to a strict class allow-list
- Terminate TLS and enforce authentication at the Pulsar broker to limit which clients can publish to consumer topics
- Avoid using getObject() on messages from untrusted sources and migrate consumers to safer serialization formats
# Configuration example: restrict Java deserialization at JVM startup
java -Djdk.serialFilter='com.yourorg.allowed.**;java.lang.*;!*' \
-Djava.security.manager \
-jar your-pulsar-consumer.jar
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB

