The vulnerability is a pre-authentication remote code execution in OpenDJ's JMX RMI connector, caused by the deserialization of untrusted data. The analysis began by identifying the patched version (5.1.1) and the last vulnerable version (5.1.0) from the advisory. By comparing the git tags for these versions, I located the specific commit that addressed the vulnerability, titled 'CVE-2026-46495 OpenDJ Unauthenticated RCE via Java Deserialization in JMX RMI'.
The commit analysis revealed two key changes. First, in RmiConnector.java, the startConnectorNoClientCertificate method was modified to call a new method, configureJmxDeserializationProtection. This new method sets a restrictive deserialization filter on the JMX environment, which is the primary fix to prevent arbitrary object deserialization. This indicates that startConnectorNoClientCertificate was responsible for setting up the insecure connector.
Second, in RmiAuthenticator.java, the authenticate method, which processes the credentials object, was hardened. Previously, it unsafely cast the input Object to an Object[]. This is the sink where the malicious deserialized object would be handled, triggering the exploit. The patch adds strict type and length checking, ensuring the credentials are a simple String[].
Therefore, org.opends.server.protocols.jmx.RmiAuthenticator.authenticate is the function that would appear in a runtime profile as it directly handles the malicious payload, and org.opends.server.protocols.jmx.RmiConnector.startConnectorNoClientCertificate is the function whose insecure setup enables the vulnerability.