Airflow

The authenticated /ui/dags endpoint did not enforce per-DAG access control on embedded Human-in-the-Loop (HITL) and TaskInstance records: a logged-in Airflow user with read access to at least one DAG could retrieve HITL prompts (including their request parameters) and full TaskInstance details for DAGs outside their authorized scope. Because HITL prompts and TaskInstance fields routinely carry operator parameters and free-form context attached to a task, the leak widens visibility of DAG-run data beyond the intended per-DAG RBAC boundary for every authenticated user. Users are recommended to upgrade to version 3.2.1 , which fixes this issue.Show less

Secrets in Variables saved as JSON dictionaries were not properly redacted - in case thee variables were retrieved by the user the secrets stored as nested fields were not masked. If you do not store variables with sensitive values in JSON form, you are not affected. Otherwise please upgrade to Apache Airflow 3.2.0 that has the fix implementedShow less

In case of SQL errors, exception/stack trace of errors was exposed in API even if "api/expose_stack_traces" was set to false. That could lead to exposing additional information to potential attacker. Users are recommended to upgrade to Apache Airflow 3.2.0, which fixes the issue.Show less

An example of BashOperator in Airflow documentation suggested a way of passing dag_run.conf in the way that could cause unsanitized user input to be used to escalate privileges of UI user to allow execute code on worker. Users should review if any of their own DAGs have adopted this incorrect advice.Show less

Dag Authors, who normally should not be able to execute code in the webserver context could craft XCom payload causing the webserver to execute arbitrary code. Since Dag Authors are already highly trusted, severity of this issue is Low. Users are recommended to upgrade to Apache Airflow 3.2.0, which fixes the issue.Show less

JWT Tokens used by tasks were exposed in logs. This could allow UI users to act as Dag Authors. Users are advised to upgrade to Airflow version that contains fix. Users are recommended to upgrade to version 3.2.0, which fixes this issue.Show less

The `access_key` and `connection_string` connection properties were not marked as sensitive names in secrets masker. This means that user with read permission could see the values in Connection UI, as well as when Connection was accidentaly logged to logs, those values could be seen in the logs. Azure Service Bus used those properties to store sensitive values. Possibly other providers could be also affected if they used the same fields to store sensitive data. If you used Azure Service Bus connection with those values set or if you have other connections with those values storing sensitve values, you should upgrade Airflow to 3.1.8Show less

The example example_xcom that was included in airflow documentation implemented unsafe pattern of reading value from xcom in the way that could be exploited to allow UI user who had access to modify XComs to perform arbitrary execution of code on the worker. Since the UI users are already highly trusted, this is a Low severity vulnerability. It does not affect Airflow release - example_dags are not supposed to be enabled in production environment, however users following the example could replicate the bad pattern. Documentation of Airflow 3.2.0 contains version of the example with improved resiliance for that case. Users who followed that pattern are advised to adjust their implementations accordingly.Show less

Dag Authors, who normally should not be able to execute code in the webserver context could craft XCom payload causing the webserver to execute arbitrary code. Since Dag Authors are already highly trusted, severity of this issue is Low. Users are recommended to upgrade to Apache Airflow 3.2.0, which resolves this issue.Show less

Before Airflow 3.2.0, it was unclear that secure Airflow deployments require the Deployment Manager to take appropriate actions and pay attention to security details and security model of Airflow. Some assumptions the Deployment Manager could make were not clear or explicit enough, even though Airflow's intentions and security model of Airflow did not suggest different assumptions. The overall security model [1], workload isolation [2], and JWT authentication details [3] are now described in more detail. Users concerned with role isolation and following the Airflow security model of Airflow are advised to upgrade to Airflow 3.2, where several security improvements have been implemented. They should also read and follow the relevant documents to make sure that their deployment is secure enough. It also clarifies that the Deployment Manager is ultimately responsible for securing your Airflow deployment. This had also been communicated via Airflow 3.2.0 Blog announcement [4]. [1] Security Model: https://airflow.apache.org/docs/apache-airflow/stable/security/jwt_token_authentication.html [2] Workload isolation: https://airflow.apache.org/docs/apache-airflow/stable/security/workload.html [3] JWT Token authentication: https://airflow.apache.org/docs/apache-airflow/stable/security/jwt_token_authentication.html [4] Airflow 3.2.0 Blog announcement: https://airflow.apache.org/blog/airflow-3.2.0/ Users are recommended to upgrade to version 3.2.0, which fixes this issue.Show less

When user logged out, the JWT token the user had authtenticated with was not invalidated, which could lead to reuse of that token in case it was intercepted. In Airflow 3.2 we implemented the mechanism that implements token invalidation at logout. Users who are concerned about the logout scenario and possibility of intercepting the tokens, should upgrade to Airflow 3.2+ Users are recommended to upgrade to version 3.2.0, which fixes this issue.Show less

Apache Airflow versions 3.0.0 through 3.1.8 DagRun wait endpoint returns XCom result values even to users who only have DAG Run read permissions, such as the Viewer role.This behavior conflicts with the FAB RBAC model, which treats XCom as a separate protected resource, and with the security model documentation that defines the Viewer role as read-only. Airflow uses the FAB Auth Manager to manage access control on a per-resource basis. The Viewer role is intended to be read-only by default, and the security model documentation defines Viewer users as those who can inspect DAGs without accessing sensitive execution results. Users are recommended to upgrade to Apache Airflow 3.2.0 which resolves this issue.Show less

Apache Airflow versions 3.1.0 through 3.1.7 missing authorization vulnerability in the Execution API's Human-in-the-Loop (HITL) endpoints that allows any authenticated task instance to read, approve, or reject HITL workflows belonging to any other task instance. Users are recommended to upgrade to Apache Airflow 3.1.8 or later, which resolves this issue.Show less

Apache Airflow versions 3.1.0 through 3.1.7 session token (_token) in cookies is set to path=/ regardless of the configured [webserver] base_url or [api] base_url. This allows any application co-hosted under the same domain to capture valid Airflow session tokens from HTTP request headers, allowing full session takeover without attacking Airflow itself. Users are recommended to upgrade to Apache Airflow 3.1.8 or later, which resolves this issue.Show less

Apache Airflow versions 3.1.0 through 3.1.7 /ui/dependencies endpoint returns the full DAG dependency graph without filtering by authorized DAG IDs. This allows an authenticated user with only DAG Dependencies permission to enumerate DAGs they are not authorized to view. Users are recommended to upgrade to Apache Airflow 3.1.8 or later, which resolves this issue.Show less

Apache Airflow versions 3.0.0 through 3.1.7 FastAPI DagVersion listing API does not apply per-DAG authorization filtering when the request is made with dag_id set to "~" (wildcard for all DAGs). As a result, version metadata of DAGs that the requester is not authorized to access is returned. Users are recommended to upgrade to Apache Airflow 3.1.8 or later, which resolves this issue.Show less

Airflow versions before 2.11.1 have a vulnerability that allows authenticated users with audit log access to see sensitive values in audit logs which they should not see. When sensitive connection parameters were set via airflow CLI, values of those variables appeared in the audit log and were stored unencrypted in the Airflow database. While this risk is limited to users with audit log access, it is recommended to upgrade to Airflow 2.11.1 or a later version, which addresses this issue. Users who previously used the CLI to set connections should manually delete entries with those connection sensitive values from the log table. This is similar but not the same issue as CVE-2024-50378Show less

DAG Author (who already has quite a lot of permissions) could manipulate database of Airflow 2 in the way to execute arbitrary code in the web-server context, which they should normally not be able to do, leading to potentially remote code execution in the context of web-server (server-side) as a result of a user viewing historical task information. The functionality responsible for that (log template history) has been disabled by default in 2.11.1 and users should upgrade to Airflow 3 if they want to continue to use log template history. They can also manually modify historical log file names if they want to see historical logs that were generated before the last log template change.Show less

When a DAG failed during parsing, Airflow’s error-reporting in the UI could include the full kwargs passed to the operators. If those kwargs contained sensitive values (such as secrets), they might be exposed in the UI tracebacks to authenticated users who had permission to view that DAG.  The issue has been fixed in Airflow 3.1.4 and 2.11.1, and users are strongly advised to upgrade to prevent potential disclosure of sensitive information.Show less