PatchSiren cyber security CVE debrief
CVE-2026-11816 keras-team CVE debrief
A path traversal vulnerability was discovered in the archive extraction utilities of Keras, specifically in the `filter_safe_tarinfos()` and `filter_safe_zipinfos()` functions located in `keras/src/utils/file_utils.py`. These functions validate archive member paths against the process's current working directory (CWD) instead of the actual extraction destination. This can be exploited when the process runs with the CWD set to `/`, a common scenario in Docker containers, CI/CD runners, and Jupyter environments, effectively making the validation boundary the filesystem root. Consequently, traversal paths can bypass the security check. Additionally, there's a bug in the zip filter that causes an `AttributeError` when encountering a blocked entry, leading to incomplete extraction. Python 3.11 installations are particularly vulnerable as they lack the `filter='data'` safety net, making them entirely dependent on the flawed CWD-based filter. Successful exploitation can lead to arbitrary file writes outside the intended extraction directory, enabling attackers to overwrite configuration files, inject malicious code, or corrupt machine learning datasets and pipelines.
- Vendor
- keras-team
- Product
- keras-team/keras
- CVSS
- HIGH 8.1
- CISA KEV
- Not listed in stored evidence
- Original CVE published
- 2026-06-11
- Original CVE updated
- 2026-07-21
- Advisory published
- 2026-06-11
- Advisory updated
- 2026-07-21
Who should care
Users of Keras versions prior to 3.14.0, especially those running in environments like Docker containers, CI/CD runners, or Jupyter environments, should be aware of this vulnerability and take steps to mitigate it.
Technical summary
The vulnerability is caused by the insecure path validation in `filter_safe_tarinfos()` and `filter_safe_zipinfos()`. The functions do not correctly validate archive member paths against the extraction destination, allowing for path traversal attacks.
Defensive priority
High
Recommended defensive actions
- Upgrade to Keras version 3.14.0 or later.
- Ensure that the CWD is properly set and validated before running archive extraction utilities.
- Use additional security measures, such as running the extraction utilities in a sandboxed environment.
Evidence notes
The vulnerability was reported by security researchers at Huntr. The CVE record and NVD details provide additional information on the vulnerability.
Sources and references
Verified primary and authoritative sources
-
CVE-2026-11816 CVE Program record
Publisher, destination, and source semantics verified
URL: https://www.cve.org/CVERecord?id=CVE-2026-11816
CVE Program - Official CVE Program record with source-provided CVE metadata.
-
CVE-2026-11816 NVD vulnerability detail
Publisher, destination, and source semantics verified
URL: https://nvd.nist.gov/vuln/detail/CVE-2026-11816
NIST National Vulnerability Database - Official NIST NVD detail page and source-specific vulnerability assessment.
Supplemental references
-
Source reference
Unverified legacy reference
URL: https://github.com/keras-team/keras/commit/2465b6657b02c8eed308759b7e800e295ae01888
-
Source reference
Unverified legacy reference
URL: https://huntr.com/bounties/a07e3983-7158-4419-af2b-38f1dea01a4f
Methodology and review provenance
AI-assisted synthesis based on stored public vulnerability evidence. System validation, approval state, and publication status do not by themselves establish human review of this revision. PatchSiren helps prioritize defensive review and does not prove exposure or remediation on any system.