Executive brief
Apache Airflow is an open-source platform used to manage and schedule complex data workflows (DAGs). A security flaw in the user interface allowed authorized users to see the names and structural details of workflows they were not supposed to access. This could lead to the exposure of private project names or internal team dependencies in environments where multiple teams share the same Airflow instance.
Technical details
An improper authorization vulnerability exists in the `/ui/structure/structure_data` endpoint of Apache Airflow. When the endpoint is called with `external_dependencies=true`, it fails to verify if the authenticated user has read permissions for linked Directed Acyclic Graphs (DAGs) before returning dependency graph nodes. An attacker with low privileges (authorized for at least one DAG) can exploit this to discover the existence and trigger/sensor structure of other DAGs they are not authorized to view. This bypasses per-DAG read scoping intended to isolate team workflows. The issue is resolved in version 3.2.2 by implementing the `ReadableDagsFilterDep` check.
Affected products
- Apache Airflow >= 3.0.0, < 3.2.2
Timeline
- 2026-05-31: disclosed: Initial disclosure on oss-security mailing list
- 2026-06-01: advisory: GitHub and NVD advisories published
- 2026-04-20: patched: Fix merged into main branch and milestone set for 3.2.2