External risk intelligence

Apache Airflow CeleryExecutor Arbitrary Code Execution via Forged Callback

CVE advisorySeverity: CRITICAL (CVSS 9.8)

CVE-2026-68982

This vulnerability affects the CeleryExecutor component within Apache Airflow, which relies on a Celery broker (like Redis or RabbitMQ) for task queuing. While network-reachable, Celery brokers are typically deployed within isolated internal networks and are not intended to be exposed directly to the public internet, making direct external exploitation unlikely in common, properly secured deployments.

Missing Authentication

Halo Surface Signal: 2 out of 5 — less likely to be public-facing.

External exposure likelihood

Horizon Alert

Summary of the vulnerability and why it matters

A security flaw in Apache Airflow's CeleryExecutor could allow unauthorized code execution on worker systems if an attacker can access the Celery broker. This means malicious commands could be run on affected systems without needing valid Airflow credentials, potentially impacting operations that rely on these workers.

Unvalidated callbacks could allow code execution.
Important for systems using CeleryExecutor.
Confirm relevance and exposure to Airflow CeleryExecutor.

Attack Path

How an attacker could exploit the issue

An attacker who can send messages to Apache Airflow's Celery broker can enqueue a malicious callback workload. This forged workload bypasses authentication and can trigger arbitrary code execution on all Airflow CeleryExecutor workers.

Entry condition: Ability to write to the Celery broker.
Trigger point: Enqueuing a forged callback workload.
Resulting risk: Arbitrary code execution on workers.

Live Threat

Current exploitation, exposure, and threat context

This vulnerability could allow an attacker to execute arbitrary code on Apache Airflow CeleryExecutor workers. This is possible when the Celery broker is accessible and the attacker can enqueue a forged callback, bypassing authentication checks. The impact is limited to deployments actively using the CeleryExecutor.

Arbitrary code execution on workers.
Enqueueing forged callbacks via broker.
Compromised worker performance and data.

Operational Fix

Recommended remediation, mitigation, and detection steps

Given the context, Apache Airflow deployments using the CeleryExecutor are at risk. The first practical step is to confirm the presence and reachability of these deployments, identify the accountable system owners—likely platform or infrastructure teams—and then prioritize remediation efforts based on business criticality and exposure.

Identify accountable platform/infrastructure owners.
Verify CeleryExecutor reachability and criticality.
Plan remediation based on confirmed risk.

Supplementary metadata

CVSS vector

Validate whether this threat affects your internet-facing exposure.

Halo Threat Intelligence helps prioritize remediation with Halo Surface Signal and H/A/L/O context. Start exposure validation with a free external attack surface trial.

Free EASM Trial See How Signal Works

Frequently asked questions

What is Apache Airflow and the CeleryExecutor?

Apache Airflow is a platform used to programmatically author, schedule, and monitor data pipelines. It uses executors to determine how tasks are run. The CeleryExecutor is a specific, scalable configuration that offloads task execution to a distributed pool of worker nodes, relying on a message broker like Redis or RabbitMQ to coordinate these tasks.

What does CVE-2026-68982 mean for system security?

This vulnerability is classified as CWE-306, which involves missing authentication for critical functions. In this case, the CeleryExecutor fails to verify the authenticity of callback workloads. An attacker can exploit this by sending a forged instruction through the message broker, causing the worker node to execute unauthorized code as if it were a legitimate task.

How is a malicious callback triggered in this CVE?

An attacker triggers the bug by directly writing a forged task message to the Celery message broker. This bypasses the need for Airflow credentials entirely. Note that if an attacker cannot reach or send messages to the Celery broker, they cannot trigger this vulnerability, regardless of their other access levels.

Why does Halo Surface Signal categorize this risk?

Halo Surface Signal flags this as 'Unlikely' for direct external exploitation because Celery brokers are typically sequestered within internal networks. While the vulnerability allows code execution, the primary defense is the architecture of the deployment; brokers are not intended to be exposed to the public internet, which limits the attack surface for remote threats.

What steps should I take if I use CeleryExecutor?

First, verify if your environment uses the CeleryExecutor configuration. Identify the infrastructure teams responsible for managing your Celery broker and worker nodes. Prioritize updating your Apache Airflow installation to version 3.4.0 or later, which introduces a secure token validation process for worker callback tasks.

References

Written by Nicholas Merritt

Nicholas Merritt is VP of Security Solutions at Halo Security. With more than 25 years across application security, vulnerability prioritization, and offensive security, he helps teams translate threat data into practical exposure validation and remediation focus. He authors Halo Threat Intelligence CVE advisories using Halo Surface Signal and H/A/L/O context to prioritize internet-facing risk.