CVE-2026-4035

Affected versions of the `mlflow/mlflow` package are vulnerable to Improper Input Validation due to the resolution of environment variables in AI Gateway secrets. The `api_key` field in gateway secrets can accept `$ENV_VAR` references, which are resolved against the MLflow server's environment during runtime, leading to the inclusion of these resolved secrets in provider authentication headers sent to the configured upstream `api_base`. An attacker can exploit this by using low-privileged authenticated access in basic-auth deployments or unauthenticated access in default deployments without `basic-auth` to exfiltrate sensitive server-side environment credentials, potentially leading to artifact poisoning and cross-boundary code execution.