{"id":"CVE-2026-67260","aliases":["BIT-airflow-2026-67260","EUVD-2026-57275","GHSA-q4c3-7575-55j2","PYSEC-2026-3707"],"status":"published","title":"Apache Airflow: DAG-author remote code execution on the Scheduler via awaiting_input next_kwargs deserialization","description":"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.","severity":{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L","level":"high","assessor":"adp","from":"cve"},"epss":{"day":"2026-11-10","score":0.00779,"percentile":0.56212,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":true},"cwes":["CWE-502"],"published":"2026-08-12T15:33:02Z","modified":"2026-11-04T02:30:58Z","fixes":[{"ecosystem":"PyPI","package":"apache-airflow","fixed_in":["3.3.1"],"sources":["ghsa","osv"]},{"ecosystem":"Bitnami","package":"airflow","fixed_in":["3.3.1"],"sources":["osv"]}],"affected":[{"source":"ghsa","ecosystem":"PyPI","package":"apache-airflow","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"3.3.0"},{"fixed":"3.3.1"}]}]},{"source":"osv","ecosystem":"Bitnami","package":"airflow","ranges":[{"type":"SEMVER","events":[{"introduced":"3.3.0"},{"fixed":"3.3.1"}]}]},{"source":"osv","ecosystem":"PyPI","package":"apache-airflow","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"3.3.0"},{"fixed":"3.3.1"}]}],"versions":["3.3.0","3.3.1rc1","3.3.1rc2"]},{"source":"osv","ecosystem":"PyPI","package":"apache-airflow","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"3.3.0"},{"fixed":"3.3.1"}]}],"versions":["3.3.0","3.3.1rc1","3.3.1rc2"]}],"products":[{"vendor":"apache","product":"airflow"},{"vendor":"apache software foundation","product":"apache airflow"}],"references":[{"url":"https://github.com/apache/airflow/pull/70685","tags":["patch","Issue Tracking","Patch","fix","web"]},{"url":"https://lists.apache.org/thread/vygr0fh82cjzjp5k4vtfmboxryvm3lyn","tags":["vendor-advisory","Mailing List","Vendor Advisory","advisory","web"]},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-58076","tags":["related","Not Applicable","web"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-67260","tags":["web","advisory"]},{"url":"https://github.com/apache/airflow/commit/18b622920cf0f486155075125ecdcabc752b1eca","tags":["web"]},{"url":"https://github.com/apache/airflow/commit/a1ab0653121829675ad864080d2cd9d3d1688b01","tags":["web"]},{"url":"https://github.com/apache/airflow/releases/tag/3.3.1","tags":["web"]},{"url":"https://github.com/pypa/advisory-database/tree/main/vulns/apache-airflow/PYSEC-2026-3707.yaml","tags":["web"]},{"url":"https://github.com/advisories/GHSA-q4c3-7575-55j2","tags":["advisory"]},{"url":"https://pypi.python.org","tags":["web"]},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/67xxx/CVE-2026-67260.json","tags":["advisory"]},{"url":"https://github.com/apache/airflow","tags":["package"]}],"sources":{"cve":[{"id":"CVE-2026-67260","title":"Apache Airflow: DAG-author remote code execution on the Scheduler via awaiting_input next_kwargs deserialization","description":"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.","status":"published","published":"2026-08-12T15:33:02.858Z","modified":"2026-11-03T06:59:09.28Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L","level":"high","assessor":"adp"}],"cwes":["CWE-502"],"updated_at":"2026-11-03T07:42:18.253008Z"}],"euvd":[{"id":"EUVD-2026-57275","description":"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.","status":"published","published":"2026-08-12T15:33:02Z","modified":"2026-11-03T06:59:09Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L","level":"high"}],"updated_at":"2026-11-03T08:50:00.381029Z"}],"ghsa":[{"id":"GHSA-q4c3-7575-55j2","title":"Apache Airflow scheduler unsafely deserializes human-in-the-loop task next_kwargs","description":"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.","status":"published","published":"2026-08-12T18:31:18Z","modified":"2026-10-02T23:15:50Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L","level":"high"}],"cwes":["CWE-502"],"updated_at":"2026-11-02T14:41:25.889593Z"}],"nvd":[{"id":"CVE-2026-67260","description":"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.","status":"published","published":"2026-08-12T16:17:14.813Z","modified":"2026-11-03T07:20:56.523Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L","level":"high","assessor":"secondary"}],"cwes":["CWE-502"],"updated_at":"2026-11-03T11:52:51.440043Z"}],"osv":[{"id":"CVE-2026-67260","title":"Apache Airflow: DAG-author remote code execution on the Scheduler via awaiting_input next_kwargs deserialization","description":"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.","status":"published","published":"2026-08-12T15:33:02.858Z","modified":"2026-11-04T02:30:58.177772559Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L","level":"high"}],"cwes":["CWE-502"],"updated_at":"2026-11-04T02:54:44.889498Z"},{"id":"GHSA-q4c3-7575-55j2","title":"Apache Airflow scheduler unsafely deserializes human-in-the-loop task next_kwargs","description":"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.","status":"published","published":"2026-08-12T18:31:18Z","modified":"2026-10-02T23:41:04.597715151Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L","level":"high"}],"cwes":["CWE-502"],"updated_at":"2026-11-02T15:46:54.055966Z"},{"id":"PYSEC-2026-3707","description":"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.","status":"published","published":"2026-08-12T16:17:14.813Z","modified":"2026-10-15T14:45:06.283288132Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L","level":"high"}],"updated_at":"2026-11-02T15:52:57.255153Z"},{"id":"BIT-airflow-2026-67260","title":"Apache Airflow: DAG-author remote code execution on the Scheduler via awaiting_input next_kwargs deserialization","description":"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.","status":"published","published":"2026-08-18T10:05:12.378Z","modified":"2026-10-02T23:41:04.597715151Z","updated_at":"2026-11-02T15:28:28.267217Z"}]}}