---
id: "GHSA-g6w4-jvgp-68x2"
kind: "vulnerability"
aliases: ["GHSA-g6w4-jvgp-68x2"]
title: "CVE-2026-110248 - Insertion of sensitive information into log files in Amazon SageMaker Python SDK"
severity: "medium"
cvss: null
epss: null
exploited: false
status: "open"
category: "package"
products: []
ecosystems: []
tags: ["aws", "lab-disclosure", "ml-pipelines", "package", "python", "sagemaker"]
not_affected_if: ["no-credentials-in-git-urls", "patched-version"]
published_at: "2026-10-14T00:00:00+00:00"
first_seen_at: "2026-10-24T20:05:36+00:00"
updated_at: "2026-10-25T05:27:51+00:00"
sources: [{"source": "lab", "url": "https://aws.amazon.com/security/security-bulletins/rss/2026-134-aws/", "first_seen_at": "2026-10-24T20:05:36+00:00"}]
---

# CVE-2026-110248 - Insertion of sensitive information into log files in Amazon SageMaker Python SDK

When a Git clone fails inside the SageMaker Python SDK, the error output includes repository credentials, which are written to CloudWatch Logs without redaction. Anyone authenticated who can read those logs can recover the plaintext credentials.

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

## How to check

- Check the installed SDK version with `pip show sagemaker` and compare against `2.257.3` (V2) or `3.10.0` (V3).
- Search CloudWatch Logs for SageMaker output from failed Git clone operations for embedded credentials in repository URLs.
- Look for Git repository URLs that contain usernames, tokens or passwords in SageMaker job or pipeline definitions.

## Fix

- Upgrade the SageMaker Python SDK V2 to `2.257.3` or V3 to `3.10.0`.
- Ensure any forked or derivative code includes the fix.
- Avoid putting credentials in Git repository URLs.
- Rotate any credentials that appeared in affected CloudWatch log output.

## Timeline

- 2026-10-24T20:05:36+00:00 lab https://aws.amazon.com/security/security-bulletins/rss/2026-134-aws/ — AWS: CVE-2026-110248 - Insertion of sensitive information into log files in Amazon SageMaker Python SDK

## Documents

- [advisory] AWS: CVE-2026-110248 - Insertion of sensitive information into log files in Amazon SageMaker Python SDK https://aws.amazon.com/security/security-bulletins/rss/2026-134-aws/

## From the sources

CVE-2026-110248 - Insertion of sensitive information into log files in Amazon SageMaker Python SDK

### CVE-2026-110248 - Insertion of sensitive information into log files in Amazon SageMaker Python SDK

Bulletin ID: 2026-134-AWS Scope: AWS Content Type: Important (requires attention) Publication Date: 10/14/2026 11:30 AM PDT

Description:

Amazon SageMaker Python SDK is an open source library for training and deploying machine learning models on Amazon SageMaker.

We identified CVE-2026-110248, an issue in the git_utils.py component of Amazon SageMaker Python SDK. When a Git clone operation fails, repository credentials in subprocess error output are written to CloudWatch Logs without redaction. An authenticated remote user with access to those logs is able to obtain the plaintext credentials.

Impacted versions:

- Amazon SageMaker Python SDK V2 versions before 2.257.3
- Amazon SageMaker Python SDK V3 versions before 3.10.0

Resolution:

This issue has been addressed in Amazon SageMaker Python SDK V2 version 2.257.3 and V3 version 3.10.0. We recommend upgrading to the latest version and ensuring any forked or derivative code incorporates the fixes.

Workarounds:

If immediate upgrading is not possible, avoid including credentials in Git repository URLs and restrict access to CloudWatch Logs containing SageMaker Python SDK output. Review logs from failed Git clone operations and rotate any credentials that were present in the affected log output.

References:

- CVE-2026-110248
- GHSA-g6w4-jvgp-68x2

Acknowledgement:

We would like to thank Gouri Sankar A for collaborating on this issue through the coordinated vulnerability disclosure process.

 Please email aws-security@amazon.com with any security questions or concerns.

---

_Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai_