CVE-2026-64849: MLflow Server-Side Request Forgery Vulnerability

An official source reports active exploitation. Teams running MLflow should verify exposure and begin risk-reduction work now.

MLflow contains a server-side request forgery vulnerability that can allow attackers to reach internal or cloud metadata services and receive response_status and response_body.

What changed

On August 20, 2026, this issue entered the urgent InfoSecNexus queue because exploitation is identified by an authoritative source. The source record, affected versions, and vendor remediation remain the controlling references; asset inventory and network context determine which systems should move first.

Why this matters

MLflow belongs in an exploit-led queue only after the affected product is matched to a reachable asset and accountable owner.

Exploit-first prioritization does not mean patching blindly. Confirm the vulnerable component is installed, identify the reachable attack path, preserve evidence of suspicious activity, and protect critical workloads while the permanent fix is deployed.

Immediate response plan

1. Identify exact product versions, owners, exposure paths, and business-critical dependencies.
2. Apply the latest vendor remediation or isolate the vulnerable path when immediate patching is not possible.
3. Review available telemetry for exploitation attempts before restarting, rebuilding, or rotating evidence away.
4. Validate the fixed version and control health, then record any exception with an owner and expiry date.