MLflow: Unauthenticated full-read SSRF in webhook delivery: _validate_webhook_url bypassed via unvalidated HTTP redirects (and DNS rebinding)
Patch now. It under exploitation confirmed by CISA and has a working public exploit.
Apply mitigations in accordance with vendor instructions, ensuring compliance with CISA’s BOD 26-04 Prioritizing Security Updates Based on Risk (see URL in Notes) guidance and CISA’s “Forensics Triage Requirements” (see URL in Notes). Follow applicable BOD 26-04 guidance for cloud services or discontinue use of the product if mitigations are unavailable. Stakeholders are responsible for evaluating each asset's internet exposure and ensuring adherence to BOD 26-04 patching guidelines.
MLflow has an unauthenticated endpoint that validates a webhook URL but then follows HTTP redirects without re-validating them, allowing attackers to bypass security checks and access internal services or cloud metadata. This lets attackers read sensitive information from systems that should be protected.
CVE-2026-64849 is a Server-Side Request Forgery (SSRF) vulnerability in MLflow's webhook test endpoint (/api/2.0/mlflow/webhooks/{id}/test). The _validate_webhook_url() function validates only the initial URL, but the delivery mechanism follows HTTP redirects and re-resolves hostnames without pinning the validated address, enabling attackers to reach internal/metadata services and exfiltrate response data. No authentication is required for exploitation.