The vulnerability in Splunk AI Toolkit versions below 5.7.4 is caused by an insecure default domain allowlist that permits outbound HTTP requests to arbitrary external domains. A low-privileged user lacking admin or power roles can exploit this to make the toolkit send requests to attacker-controlled servers, potentially leading to data exfiltration. The issue stems from a default resource setting intended to be changed by administrators but left insecure by default.

An attacker with low-privileged Splunk user access can leverage this vulnerability to cause the Splunk AI Toolkit to send outbound HTTP requests to malicious servers, which may result in unauthorized data exfiltration. The confidentiality of data is impacted, but integrity and availability are not affected according to the CVSS vector.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until an official fix is available, administrators should review and restrict the domain allowlist configuration in the Splunk AI Toolkit to prevent outbound requests to unapproved domains. Monitoring and limiting user permissions to prevent exploitation may also help mitigate risk.