Apache Airflow’s /assets/events API returned asset events for every Dag in the deployment without applying a per‑Dag authorization filter. As a result, any authenticated user who holds asset‑read access could enumerate asset events—including the source Dag ID, task ID, run ID and event timestamps—for Dags they are not permitted to see. The absence of the filter in the count query also revealed the existence of hidden Dags through pagination metadata. This is a classic case of unauthorized information disclosure (CWE-200) that exposes structural details of the Airflow environment to privileged but otherwise limited users.

The vulnerability affects Apache Airflow installations that use per‑Dag access control to segregate teams or tenants. The issue exists whenever per‑Dag authorization is enabled; no special configuration beyond standard Airflow deployments is required. No affected version information is provided in the available data. The affected product is Apache Airflow from the Apache Software Foundation.

The threat vector is a remote authenticated request to the Airflow web API. An attacker needs only valid credentials with asset‑read permission to trigger the flaw and gain visibility into all Dag assets across the deployment. The CVSS score of 4.3 indicates low overall severity; EPSS data is not available and the vulnerability is not listed in CISA’s KEV, yet the ability to enumerate internal DAG information can aid further attacks such as targeted credential harvesting or knowledge of workflow structure. The absence of an authorization gate and the exposure of discovery metadata increase the risk of misuse if credential leakage occurs. While the public exploitation likelihood remains uncertain, the internal impact for organizations with segregated Airflow tenants is high.