Home › Intelligence › Brief
VULNERABILITY BRIEF🟡 Medium Vulnerability

Privilege Escalation in MLflow Tracking Server (CVE-2026-118625) Allows Authenticated Attackers to Gain Elevated Access

A newly disclosed CVE‑2026‑118625 in MLflow’s Tracking Server lets authenticated users bypass authorization and elevate privileges. The flaw affects all pre‑3.16.0 releases and is fixed in version 3.16.0. For compliance teams, the issue highlights the need for robust access‑control evidence and continuous verification of authorization logic.

LiveThreat™ Intelligence · 📅 October 30, 2026· 📰 zerodayinitiative.com
🟡
Severity
Medium
VU
Type
Vulnerability
🎯
Confidence
High
🏢
Affected
1 sector(s)
✅
Actions
3 recommended
📰
Source
zerodayinitiative.com

Privilege Escalation in MLflow Tracking Server (CVE‑2026‑118625) Allows Authenticated Attackers to Gain Elevated Access

What It Is — The Zero Day Initiative advisory (ZDI‑26‑841) details a missing authorization check in the get_endpoints function of the MLflow Tracking Server. An authenticated remote attacker can invoke this function to bypass intended access controls and elevate privileges to resources normally restricted. The flaw is assigned CVE‑2026‑118625, carries a CVSS 6.3 (Network, Low complexity, Low privileges required, No user interaction, Confidentiality Low, Integrity Low, Availability Low) and is fixed in MLflow 3.16.0. No public exploit code has been released, but the vulnerability is exploitable over the network once a valid user session is established.

Affected Products — MLflow (open‑source tracking server) – all versions prior to 3.16.0.

Why It Matters for Trust & Control Assurance — The issue demonstrates a gap in the access‑control control objective; organizations must prove that every privileged operation is gated by proper authorization checks to satisfy frameworks such as NIST CSF 2.0.

Recommended Actions —

  • Upgrade all MLflow Tracking Server instances to 3.16.0 or later.
  • Review code and configuration for any other endpoints lacking authorization checks.
  • Conduct an internal access‑control audit and capture evidence of remediation for audit readiness.

Source: ZDI Advisory

📰 Original Source
http://www.zerodayinitiative.com/advisories/ZDI-26-841/ ↗

This LiveThreat Intelligence Brief is an independent analysis. Read the original reporting at the link above.

From the Verisq platform · Trust Operations

Access is where most audits get tested.

Verisq AI Trust Operations maps incidents like this to your access controls and collects the evidence continuously, keeping your trust posture defensible.

See where you'd stand with Verisq AI Trust Operations →