CVE-2026-67260: Apache Airflow scheduler unsafely deserializes human-in-the-loop task next_kwargs
Apache Airflow 3.3.0 moved human-in-the-loop tasks from the triggerer to a new awaitinginput task state swept by the scheduler. That sweep deserializes the task instance's nextkwargs without an allow-list, so a Dag author — who controls that value through the task execution API — can cause an arbitrary module import and object instantiation inside the scheduler process, or terminate the scheduler job.
Published Aug 12, 2026 · Updated Oct 2, 2026
Affected packages
apache-airflow
- Ecosystem
- PyPI
- Affected
- >= 3.3.0, < 3.3.1
- Fixed in
- 3.3.1
- Check
pip show apache-airflow
Advisory details
From the GitHub Advisory Database (CC BY 4.0).
Apache Airflow 3.3.0 moved human-in-the-loop tasks from the triggerer to a new awaiting_input task state swept by the scheduler. That sweep deserializes the task instance's next_kwargs without an allow-list, so a Dag author — who controls that value through the task execution API — can cause an arbitrary module import and object instantiation inside the scheduler process, or terminate the scheduler job. No non-default configuration is required: the sweep runs unconditionally every 15 seconds, and the default allowed_deserialization_classes setting does not cover this code path. Versions before 3.3.0 are not affected, because human-in-the-loop tasks deferred onto the triggerer instead. This is a different code path from CVE-2026-58076, which covers the same unguarded exception-node deserialization reached elsewhere — deployments that applied that fix must upgrade for this issue as well. Users are advised to upgrade to apache-airflow 3.3.1 or later.
Credited to oscerd.
References
- https://nvd.nist.gov/vuln/detail/CVE-2026-67260
- https://github.com/apache/airflow/pull/70685
- https://lists.apache.org/thread/vygr0fh82cjzjp5k4vtfmboxryvm3lyn
- https://www.cve.org/CVERecord?id=CVE-2026-58076
- https://github.com/apache/airflow/commit/18b622920cf0f486155075125ecdcabc752b1eca
- https://github.com/apache/airflow/commit/a1ab0653121829675ad864080d2cd9d3d1688b01
- https://github.com/apache/airflow/releases/tag/3.3.1
- https://github.com/pypa/advisory-database/tree/main/vulns/apache-airflow/PYSEC-2026-3707.yaml
- https://github.com/advisories/GHSA-q4c3-7575-55j2
How this page is made: the data panel comes from the GitHub Advisory Database, CISA’s KEV catalog, the NVD and FIRST EPSS, and updates twice a day. No analysis has been written for this entry yet. Methodology
More PyPI vulnerabilities
All PyPI entriesCVE-2026-54183
Apache Airflow Variables UI fails to mask sensitive values in deeply nested iterables
medium4.3Aug 12, 2026CVE-2026-58076
Apache Airflow exception-node deserialization permits arbitrary callable execution
high8.8Aug 12, 2026CVE-2026-59242
Apache Airflow XCom API permits unsafe deserialization through JSON string literals
medium5.4Aug 12, 2026CVE-2026-59244
Apache Airflow exposes dict-valued var.json secrets in Rendered Templates
medium6.5Aug 12, 2026CVE-2026-65017
Apache Airflow Config API exposes team-scoped sensitive configuration values
medium6.5Aug 12, 2026CVE-2026-67587
Apache Airflow Task SDK Callback deserialization can import arbitrary modules in the scheduler
high8.8Aug 12, 2026
