By manipulating the webhook configuration, an attacker can force the Label Studio server to perform outbound requests on their behalf. This can be weaponized to interact with internal APIs that lack authentication or to exfiltrate sensitive environment configuration data and annotation datasets. The impact is significant for organizations hosting Label Studio within sensitive internal network segments or cloud environments where internal service discovery relies on metadata endpoints.

Impact

Successful exploitation allows authenticated attackers to bypass network perimeters, potentially leading to unauthorized data exfiltration of sensitive annotations, reconnaissance of internal network infrastructure, and compromise of internal services accessible from the Label Studio host.

Recommendation

Prioritize the identification and patching of Label Studio instances running version 1.23.0 or earlier. Audit current webhook configurations to identify any entries pointing to internal IP addresses or private domain namespaces.