AZL-111398: CVE-2026-116137 affecting package python-tensorboard 2.16.2-6

Q: When is an application using this package exposed?

Exposure depends on the application processing attacker-controlled ZIP archives with adm-zip and using validation such as readEntries(), getEntry(), or readAsText() before extracting archives with extractAllTo(). The issue relies on ZIP entries whose names are canonically equivalent, such as './hello.txt' and 'hello.txt'.

Q: What does an attacker need to exploit the issue?

An attacker needs to provide a crafted ZIP archive containing duplicate paths represented by canonically equivalent entry names. This can cause validation to inspect benign content while extraction writes malicious content to the same destination path.

Q: How can teams reduce risk if they cannot patch immediately?

Do not process untrusted ZIP archives through the affected validation-and-extraction flow. Before extraction, reject archives containing entry names that normalize to the same path, including variants such as './hello.txt' and 'hello.txt'.

Published Oct 25, 2026

Updated

adm-zip through 0.6.1 contains a duplicate-entry validation bypass that allows attackers to evade the readEntries() check by using canonically equivalent entry names like './hello.txt' and 'hello.txt'. Attackers can craft ZIP archives where getEntry() or readAsText() validates benign content while extractAllTo() writes malicious content to the same path, bypassing integrity checks.

Affected Software

1 affected component

azure-linux/python-tensorboard<=2.16.2-6

Event History

Oct 25, 2026

Advisory Published

via OSV·12:16 PM

Data Sourced

via OSV·12:16 PM

DescriptionAffected Software

Child vulnerabilities

Contains the following vulnerabilities.

CVE-2026-116137

Frequently Asked Questions

When is an application using this package exposed?

Exposure depends on the application processing attacker-controlled ZIP archives with adm-zip and using validation such as readEntries(), getEntry(), or readAsText() before extracting archives with extractAllTo(). The issue relies on ZIP entries whose names are canonically equivalent, such as './hello.txt' and 'hello.txt'.

What does an attacker need to exploit the issue?

An attacker needs to provide a crafted ZIP archive containing duplicate paths represented by canonically equivalent entry names. This can cause validation to inspect benign content while extraction writes malicious content to the same destination path.

How can teams reduce risk if they cannot patch immediately?

Do not process untrusted ZIP archives through the affected validation-and-extraction flow. Before extraction, reject archives containing entry names that normalize to the same path, including variants such as './hello.txt' and 'hello.txt'.

AZL-111398: CVE-2026-116137 affecting package python-tensorboard 2.16.2-6

Affected Software

Event History

Child vulnerabilities

Frequently Asked Questions

When is an application using this package exposed?

What does an attacker need to exploit the issue?

How can teams reduce risk if they cannot patch immediately?

Company

Resources

Contact

Frequently Asked Questions

When is an application using this package exposed?

What does an attacker need to exploit the issue?

How can teams reduce risk if they cannot patch immediately?