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

> Unauthenticated attackers can exploit Server-Side Request Forgery in MLflow webhooks to access internal services and cloud metadata via redirect bypasses.

- **Severity:** Critical 9.3 (`CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:L/A:N`)
- **Exploited in the wild:** Yes, since 19 Aug 2026 (CISA KEV)
- **Login needed:** Not needed
- **Attackable:** Over the internet
- **Victim action needed:** Not needed
- **Fix available:** Yes
- **Published:** 17 Aug 2026
- **Updated:** 2 Oct 2026
- **Also known as:** BIT-mlflow-2026-64849, GHSA-7gwp-5pfp-969j, MINI-8g3x-hmjq-4q5r, PYSEC-2026-3687

## What is CVE-2026-64849?

CVE-2026-64849 is a critical server-side request forgery vulnerability in Lfprojects Mlflow, published on 17 Aug 2026. An attacker can exploit it over the internet without logging in, with no action from the victim.

### Official description

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.

Weakness: CWE-918 (Server-Side Request Forgery)

## Which versions are affected by CVE-2026-64849?

4 open source packages ship an affected version, including mlflow. It affects 1 product.


| Package | Ecosystem | Affected | Fixed in |
| --- | --- | --- | --- |
| `mlflow` | Bitnami | From 3.3.0 | 3.15.0 |
| `mlflow` | PyPI | All versions | 3.15.0 |
| `mlflow` | PyPI | From 3.3.0 | 3.15.0 |
| `mlflow` | PyPI | From 3.3.0 | 3.15.0 |

Products: Lfprojects Mlflow.

## How do I fix CVE-2026-64849?

Upgrade mlflow to 3.15.0 or later. US federal agencies must fix it by 2 Sep 2026.

## Is CVE-2026-64849 being exploited?

Yes. Attackers have exploited CVE-2026-64849 in the wild since at least 19 Aug 2026.

Public exploit code:

- https://github.com/codeb0ssx/CVE-2026-64849-PoC (github_poc): https://github.com/codeb0ssx/CVE-2026-64849-PoC
- CVE-2026-64849 PoC (github_poc): https://github.com/BiuTrap/CVE-2026-64849
- https://github.com/zavisco/CVE-2026-64849.yaml (github_poc): https://github.com/zavisco/CVE-2026-64849.yaml
- este laboratorio puede estar bien o mal preguntale a la IA estoy probando pero debe funcionar hahahah (github_poc): https://github.com/isaca0315/CVE-2026-64849-poc-lab
- CVE-2026-64849 — MLflow SSRF (github_poc): https://github.com/maxprog-svg/CVE-2026-64849
- MLflow Webhook SSRF - Unauthenticated Full-Read via Redirect Bypass (nuclei): http/cves/2026/CVE-2026-64849.yaml

## Other vulnerabilities in Lfprojects Mlflow

- [CVE-2025-15379](https://insights.sbomify.com/CVE-2025-15379/): Command Injection in mlflow/mlflow (critical 10.0)
- [CVE-2024-0520](https://insights.sbomify.com/CVE-2024-0520/): Remote Code Execution due to Full Controlled File Write in mlflow/mlflow (critical 10.0)
- [CVE-2023-6977](https://insights.sbomify.com/CVE-2023-6977/): Path Traversal: '\..\filename' (critical 10.0)
- [CVE-2023-6709](https://insights.sbomify.com/CVE-2023-6709/): Improper Neutralization of Special Elements Used in a Template Engine in mlflow/mlflow (critical 10.0)
- [CVE-2023-6015](https://insights.sbomify.com/CVE-2023-6015/): MLflow Arbitrary File Upload (critical 10.0)

## Fix prompt for AI coding agents

```text
Check this codebase for CVE-2026-64849 and fix it if it is affected.

What it is: CVE-2026-64849 is a critical server-side request forgery vulnerability in Lfprojects Mlflow, published on 17 Aug 2026. An attacker can exploit it over the internet without logging in, with no action from the victim.
Exploitation: Yes. Attackers have exploited CVE-2026-64849 in the wild since at least 19 Aug 2026.

Affected packages and fixed versions:
- mlflow (Bitnami): affected from 3.3.0; fixed in 3.15.0
- mlflow (PyPI): affected all versions; fixed in 3.15.0
- mlflow (PyPI): affected from 3.3.0; fixed in 3.15.0
- mlflow (PyPI): affected from 3.3.0; fixed in 3.15.0

Task:
1. Search the dependency manifests and lockfiles (for example package.json, package-lock.json, requirements.txt, poetry.lock, uv.lock, pom.xml, build.gradle, go.mod, Cargo.lock, Gemfile.lock) and any Dockerfiles for the affected packages, including transitive dependencies.
2. If an affected version is in use, upgrade it to the fixed version for its release line with the project's package manager, so the lockfile is regenerated. Prefer the smallest upgrade that fixes it.
3. If there is no fix, or the upgrade would break the build, do not guess: explain the risk and the safest workaround instead.
4. Run the tests and summarise what changed and why.
5. If nothing in the project is affected, say so and change nothing.

Reference: https://insights.sbomify.com/CVE-2026-64849/ (Markdown: https://insights.sbomify.com/CVE-2026-64849/index.md)
```

## Sources

- CVE record: https://www.cve.org/CVERecord?id=CVE-2026-64849
- NVD: https://nvd.nist.gov/vuln/detail/CVE-2026-64849
- OSV: https://osv.dev/vulnerability/CVE-2026-64849

---
From sbomify Insights: https://insights.sbomify.com/CVE-2026-64849/
Summaries are generated and may contain mistakes; the official description above is the record.