---
id: "CVE-2026-127202"
page: "https://infrastack.ai/live/CVE-2026-127202"
kind: "vulnerability"
aliases: ["CVE-2026-127202"]
title: "MLflow: unauthenticated path traversal writes files anywhere on the server"
severity: "medium"
cvss: 6.3
epss: null
exploited: false
status: "open"
category: "ai-tooling"
products: []
ecosystems: []
tags: ["ai-tooling", "mlflow", "python", "self-hosted"]
not_affected_if: ["database-backend", "file-store-disabled", "managed-hosting"]
published_at: "2026-11-12T17:17:43+00:00"
first_seen_at: "2026-11-12T17:39:39+00:00"
updated_at: "2026-11-12T17:40:01+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-127202", "first_seen_at": "2026-11-12T17:39:39+00:00"}]
exploit_sources: []
---

# MLflow: unauthenticated path traversal writes files anywhere on the server

Link for people: https://infrastack.ai/live/CVE-2026-127202 (this .md is for agents)

MLflow before 3.18.0 with the FileStore backend lets an unauthenticated remote attacker supply a crafted model_id to the log outputs API. The attacker can then create directories and write a meta.yaml file at arbitrary locations the server account can reach.

_Summary, check and fix written by a model from the sources; labelled as such._

## How to check

- Check the MLflow version with `mlflow --version` or `pip show mlflow`, and treat anything below `3.18.0` as exposed.
- Check whether FileStore is in use: FileStore was the default before `3.7.0` and is opt-in afterward through `MLFLOW_ALLOW_FILE_STORE`.
- Check whether the tracking server is reachable from the network, since the endpoint is `POST /api/3.0/mlflow/runs/outputs`.

## Fix

- MLflow: move to `3.18.0` or later.

## Timeline

- 2026-11-12T17:39:39+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-127202

## Documents

- [advisory] Third-party advisory: www.vulncheck.com https://www.vulncheck.com/advisories/mlflow-path-traversal-via-model-id-in-log-outputs-api
- [fix] Patch: github.com https://github.com/mlflow/mlflow/commit/ab5fc795eedc22348253cf687818cf2a16c15282
- [fix] Release notes: github.com https://github.com/mlflow/mlflow/releases/tag/v3.18.0

## From the sources

MLflow before 3.18.0 using the FileStore backend contains a path traversal vulnerability that allows unauthenticated remote attackers to write files outside the intended run directory by supplying a malicious model_id parameter to the log outputs API. Attackers can exploit the unvalidated model_id containing ../ sequences via the POST /api/3.0/mlflow/runs/outputs endpoint to create directories and write a meta.yaml file at arbitrary filesystem locations, limited only by the server account's permissions. FileStore was the default backend before 3.7.0 and is opt-in afterward through MLFLOW_ALLOW_FILE_STORE.

---

_Free to read, with fair use of about 60 pages a minute. For higher limits or API access: sales@infrastack.ai_