CVE-2026-116691
Received Received - Intake

MLflow CrewAI Autolog Credential Exposure

Publication date: 2026-10-27

Last updated on: 2026-10-27

Assigner: GitHub, Inc.

Description
MLflow is an open source AI engineering platform for agents, large language models, and machine learning models. From 2.19.0 until 3.17.0, mlflow.crewai.autolog() records CrewAI Agent objects without consistently sanitizing nested credential-bearing values. mlflow/crewai/autolog.py::_get_agent_attributes() converts the Agent llm field to a string that includes the provider API key, while mlflow/crewai/autolog.py::_construct_full_inputs() expands the Agent object into mlflow.spanInputs. Applications can therefore send plaintext provider credentials to the MLflow trace backend or a configured OTLP destination, where trace readers can obtain them. This issue is fixed in version 3.17.0.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published
2026-10-27
Last Modified
2026-10-27
Generated
2026-10-27
AI Q&A
2026-10-27
EPSS Evaluated
N/A
NVD
CVE-2026-116691
EUVD
EUVD-2026-111381

Affected Vendors & Products

Showing 1 associated CPE
Vendor Product Version / Range
mlflow mlflow >= 2.19.0, < 3.17.0

Helpful Resources

Exploitability

CWE
CWE Icon
KEV
KEV Icon
CWE ID Description
CWE-532 The product writes sensitive information to a log file.
CWE-312 The product stores sensitive information in cleartext within a resource that might be accessible to another control sphere.
CWE-200 The product exposes sensitive information to an actor that is not explicitly authorized to have access to that information.

Attack-Flow Graph

AI Quick Actions

Instant insights powered by AI
Executive Summary

The vulnerability in MLflow versions 2.19.0 to 3.17.0 involves the mlflow.crewai.autolog() function recording CrewAI Agent objects without properly sanitizing nested credential-bearing values. Specifically, the Agent's llm field containing provider API keys is converted to a string and included in trace data sent to MLflow trace backend or OTLP destinations. This exposes plaintext credentials to trace readers.

Detection Guidance

Check MLflow versions between 2.19.0 and 3.17.0 for CrewAI Agent autologging. Inspect logs for plaintext API keys in traces sent to MLflow or OTLP destinations. Look for sensitive data in spanInputs or trace backends.

Impact Analysis

If you use MLflow versions 2.19.0 to 3.17.0 with CrewAI integration, attacker with access to trace data could extract plaintext API keys from the Agent objects. This could lead to unauthorized access to AI services, data breaches, or misuse of credentials for malicious purposes.

Compliance Impact

This vulnerability could violate GDPR and HIPAA by exposing sensitive credentials and potentially personal data in plaintext. GDPR requires protection of personal data, while HIPAA mandates safeguarding protected health information. Unauthorized access to credentials may also constitute a compliance failure under these regulations.

Mitigation Strategies

Upgrade MLflow to version 3.17.0 or later to fix the issue. Disable mlflow.crewai.autolog() if not required. Review and sanitize any stored traces containing credentials.

Chat Assistant

Ask questions about this CVE
Hi! I’m here to help you understand CVE-2026-116691. Ask me anything about the vulnerability, its impact, or mitigation strategies.
0/70

EPSS Chart