GPU VulnDB

Database/Control plane, storage & DevOps

Apache Airflow: Import Errors API shows raw DAG parse stack traces to users who should not see that file

CVSS 8.2CVE-2026-40913Control plane, storage & DevOpscurated

Impact

When a DAG file fails to parse and has no registered DAG, Airflow's public Import Errors API returns the raw stack trace to any user with ImportError.can_read. Normally the DAG's own permissions would limit who sees it. The trace can expose file paths and pieces of pipeline code, so on a shared Airflow that runs ML pipelines, one team can read details of another team's DAG files. The record describes information disclosure only, with no code execution.

Who can reach it

Authenticated Airflow user holding the ImportError.can_read permission, with network access to the Airflow API.

What to do

Upgrade to Apache Airflow 3.4.0 or later, which limits these stack traces to the new IMPORT_ERRORS_ALL permission. Then restart the Airflow API/webserver components. After upgrading, check which roles get IMPORT_ERRORS_ALL. If you cannot upgrade yet, remove ImportError.can_read from roles that should not see every DAG file.

References

Related entries

All Control plane, storage & DevOps entries

This entry is curated: imported from vendor advisories with machine assistance, not yet individually verified. Confirm against your vendor's advisory before acting, and report anything wrong.