CVE-2026-64849 in lfprojects mlflow
LogPulse has a published detection rule for this CVE: it raises a notable when someone tries to exploit it.
What the vulnerability is
MLflow is an open source AI engineering platform for agents, large language models, and machine learning models. Starting in 3.3.0 and prior to 3.15.0, the unauthenticated POST /api/2.0/mlflow/webhooks/{id}/test endpoint calls _validate_webhook_url() in mlflow/utils/validation.py only for the original URL while mlflow/webhooks/delivery.py follows redirects and re-resolves the hostname without pinning the validated address, allowing attackers to reach internal or cloud metadata services and receive response_status and response_body. This issue is fixed in version 3.15.0.
Affected software: lfprojects mlflow
Scores
- CVSS: 9.3 (critical)
- EPSS: 10.02% chance of exploitation in the next 30 days
- CISA KEV: Known exploited, listed 2026-08-19
- Published: 2026-08-17
What LogPulse has for it
- Detection rule: Yes, published
- Agent check (vulnerable version): No
- Mitigation: No
Source data: NIST NVD, CISA KEV and FIRST EPSS. LogPulse processes new CVEs every hour.
We build every detection with great care, but no two environments are the same. Log sources, parsing and configuration differ, so test a detection in your own environment before you rely on it.