CVE-2026-96775Public exploitNetwork attack vectorHinoki check available

Code execution via ungated DSPy pickle loading

Published Sep 23, 2026 · Updated Sep 23, 2026

Unsafe deserialization in MLflow 2.0 and later allows remote attackers to execute code when a crafted DSPy model artifact is loaded. In mlflow/dspy/load.py, the non-.pkl branch calls dspy.load with allow_pickle enabled without applying the disabled-deserialization control. The attacker must place the artifact in a model source, and code runs when an operator or application loads it while MLFLOW_ALLOW_PICKLE_DESERIALIZATION is false.

CVSS severityUnavailable
Unscored
EPSS probability0.39%
Next 30 days · Oct 19, 2026
Known exploitationUnconfirmed
Based on sourced intelligence
Hinoki checkAvailable
Coverage for this vulnerability

See if you're affected

Explore vulnerability checks for your environment with Hinoki.

Get started

Summary

What happened

Unsafe deserialization in MLflow 2.0 and later allows remote attackers to execute code when a crafted DSPy model artifact is loaded. In mlflow/dspy/load.py, the non-.pkl branch calls dspy.load with allow_pickle enabled without applying the disabled-deserialization control. The attacker must place the artifact in a model source, and code runs when an operator or application loads it while MLFLOW_ALLOW_PICKLE_DESERIALIZATION is false.

The record

CVE
CVE-2026-96775
Published
Sep 23, 2026
Updated
Sep 23, 2026
Vendor
MLflow Project
Product
MLflow
Classifications
T1204.002
Attack vector
network
Privileges
Unavailable

Timeline

How it unfolded

  1. Sep 23, 2026CVE publishedPublication date reported by the CVE source.
  2. Sep 23, 2026Record updatedLatest update available in the CVE record.

Exploitability

Present is not the same as exploitable

Compare your product and version with the public record. A matching version still requires validation against your environment.

Is a vulnerable build present?

Compare these published version ranges with your installed build and any vendor patches.

  1. Affected versionversion=2.0

What conditions does exploitation require?

Attack vectornetwork
Required privilegesUnavailable in this record.

What is affected?

MLflow Project · MLflowversion=2.0

Published CVSS scores

No CVSS assessment is available in this record.

CVSS describes severity. EPSS estimates exploitation probability.

Attacks

What attackers are doing with it

Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.

Activity data is delayed. The latest report is Sep 28, 2026.

Daily unique IPsNo honeypot observations are available for this CVE in the selected window.

No observations available

Sep 22, 2026Sep 28, 2026
Latest reporting daySep 28, 2026
Latest daily unique IPsUnavailable
Prior 30-day averageUnavailable
SourceShadowserver honeypots (KEV)
Vectornetwork
PrivilegesUnavailable
Known exploitationUnconfirmed
Public exploitPublished

Weakness, pattern, technique

T1204.002User Execution: Malicious File

Public exploit references

Labels summarize the accepted research assessment. They do not indicate a test against your environment.

Technologies

Your stack

See the directory against your own environment.

Your stack

Check the software in your environment

Get started to see how Hinoki identifies affected software and validates exploitability in your environment.

Get started