# CVE-2026-2614: Arbitrary File Read via Prompt Tag Source Validation Bypass in mlflow/mlflow

- Exploited (CISA KEV): no
- Published: 2026-05-11
- CVSS: 7.5 HIGH (v3.0)
- EPSS: 3.7% (90th percentile)
- CWE: CWE-22

A vulnerability in the `_create_model_version()` handler of `mlflow/server/handlers.py` in mlflow/mlflow versions 3.9.0 and earlier allows an unauthenticated remote attacker to read arbitrary files from the server's filesystem. The issue arises when a `CreateModelVersion` request includes the tag `mlflow.prompt.is_prompt`, which bypasses source path validation. This enables an attacker to store an arbitrary local filesystem path as the model version source. The `get_model_version_artifact_handler()` function later uses this source to serve files without verifying the model version's prompt status, leading to a complete confidentiality compromise. This issue is fixed in version 3.10.0.

## Edge WAF coverage (claimed by vendors, not measured)

| Vendor | Status | First rule | Days after publication |
|---|---|---|---|
| Cloudflare | No public claim | — | — |
| AWS | No public claim | — | — |
| Microsoft Azure | No public claim | — | — |
| Google Cloud | No public claim | — | — |
| Fastly | Before feed history | — | — |
| F5 | Rule published | 2026-09-16 | 128 |
| Akamai | No public source | — | — |
| OWASP CRS | No public claim | — | — |

### F5

- 2026-09-16 · new · rule 200007118
  > Added: signature 200007118 "MLflow Arbitrary File Read" (Path Traversal, risk Medium), systems: Python. References: CVE-2026-2614.
  Source: https://docs.cloud.f5.com/docs-v2/platform/changelogs/security-details-changelog

## Our test (measured)

- OWASP Core Rule Set 4.25.0, paranoia level 1, anomaly threshold 5, default CRS settings: passed (0 of 3 requests blocked), template CVE-2026-2614, 2026-11-02
- OWASP Core Rule Set 4.25.0, paranoia level 2, anomaly threshold 5, default CRS settings: partial (2 of 3 requests blocked), template CVE-2026-2614, 2026-11-02

"No public claim" means no vendor statement names this CVE; "Before feed history" means the CVE is older than the vendor feed. In both cases generic rules may still block it.

HTML: https://cve.vigilbase.com/cve/CVE-2026-2614 · JSON: https://cve.vigilbase.com/api/v1/cves/CVE-2026-2614