What is CVE-2026-64849?
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.
The record classifies CVE-2026-64849 under weakness class CWE-918. Its CVSS v3 base vector states that the flaw is reachable remotely over the network, needs no prior authentication, needs no user interaction, and has high impact on confidentiality. 1 affected-product entry is recorded, across 1 vendor, listed below. The identifier was first published 72 days ago.
Severity and exploitation probability
- CVSS v3.1 base score
- 9.3 — CRITICAL
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:L/A:N - EPSS (FIRST)
- 9.8% probability of exploitation in the next 30 days, higher than 95.4% of all scored CVEs
- CISA KEV
- Listed since 2026-08-19, federal remediation deadline 2026-09-02
- Threadlinqs priority
- 9/10 — CISA KEV-listed, which Threadlinqs floors at 9
- Published
- 2026-08-17, last modified 2026-10-02
Is CVE-2026-64849 being exploited?
CISA added CVE-2026-64849 to the Known Exploited Vulnerabilities catalog on 2026-08-19, which means the agency holds evidence of exploitation in the wild; US federal civilian agencies had to remediate it by 2026-09-02 under BOD 22-01. Weaponised exploit code for CVE-2026-64849 is publicly available. 5 public proof-of-concept repositories are tracked for this identifier. A ProjectDiscovery Nuclei detection template exists (http/cves/2026/CVE-2026-64849.yaml). It currently carries a trending score of 65 in the Threadlinqs vulnerability feed.
- codeb0ssx/CVE-2026-64849-PoC (github)
- BiuTrap/CVE-2026-64849 (github)
- zavisco/CVE-2026-64849.yaml (github)
- isaca0315/CVE-2026-64849-poc-lab (github)
- trickest/cve (trickest)
Affected products and versions
- mlflow: mlflow
Affected software packages:
pip:mlflow— patched in 3.15.0
How to fix CVE-2026-64849
The record marks a vendor fix as available for CVE-2026-64849. Fixed versions are recorded for 1 package: pip:mlflow 3.15.0. Patch reference: https://github.com/mlflow/mlflow/commit/ba949522477cbd5915aa55d29b0cfad7d5ddf939. Vendor advisory: https://github.com/mlflow/mlflow/security/advisories/GHSA-7gwp-5pfp-969j. Because CVE-2026-64849 is KEV-listed, US federal civilian agencies were required to apply the vendor fix, or stop using the product, by 2026-09-02. Apply the vendor fix referenced above to every affected product listed in this record, then confirm the running version against the vendor advisory.
Threat activity tracking CVE-2026-64849
1 tracked threat in the Threadlinqs corpus references CVE-2026-64849, either in the campaign’s CVE list or as an indicator on the campaign record.
Sources
Enriched from CVE.org, NVD, FIRST EPSS, CISA KEV, GitHub Security Advisories, public proof-of-concept repositories, ProjectDiscovery Nuclei. Last verified by Threadlinqs on . This product uses the NVD API but is not endorsed or certified by the NVD.