MediumVulnerability

CVE-2026-44723: Vowpal Wabbit command execution through pull request title in GitHub workflow

Source: NVD(opens in a new tab)
Identifier: CVE-2026-44723
Published: May 26, 2026
Record updated: October 9, 2026

Known exploitation: Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
EPSS: 0.6%

Summary

CVE-2026-44723 affects Vowpal Wabbit, a machine learning system. The workflow .github/workflows/python_checks.yml inserts ${{ github.event.pull_request.title }} directly into double-quoted bash strings in four steps across four jobs, passing it to run_tests_model_gen_and_load.py. Because the shell expands the title before Python runs, an attacker can break out of the quotes and run arbitrary commands on the runner, and the pull_request trigger fires for PRs targeting any branch with no further access gate.

Mitigation

This vulnerability is fixed by the 998e390e80a7e8192d7849b7784bc113dbd190ad commit.

Details

Scores and exploitation

CVSS score: 5(medium)
CVSS vector: CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:L/I:N/A:N
Attack vector: network
Attack complexity: low
Privileges required: low
User interaction: none
EPSS: 0.6%Estimated probability that this CVE is exploited in the next 30 days, from the FIRST Exploit Prediction Scoring System. Read on October 11, 2026.
Disclosure date: May 26, 2026

Weaknesses and techniques

Weakness (CWE): CWE-78
CWE-1336
Attack pattern (CAPEC): CAPEC-88
AI threat technique (MITRE ATLAS): AML.T0010

Classification

Attack type: Supply chain
Attack sophistication: Trivial
AI component targeted: Framework
Impact: Integrity,
Confidentiality
Labels: security,
industry

How this record was produced

Summaries, mitigations and labels on this page are generated automatically from the source. The linked source is authoritative. The source gives this advisory a description and no title, so the title above is a headline generated from the source text.

Original source: https://nvd.nist.gov/vuln/detail/CVE-2026-44723

First tracked: May 26, 2026 at 08:12 PM ET, from NVD/CVE Database

Classifier: claude-haiku-5-5, prompt v4

Classifier confidence: 80% (self-assessed)

Summary

Mitigation

Related items