Path Traversal in Apache Airflow Google Provider

Executive Summary

CVE-2026-49297 is a security vulnerability in Apache Airflow's Google Cloud Storage operators, specifically GCSToSFTPOperator and GCSTimeSpanFileTransformOperator. These operators join Google Cloud Storage (GCS) object names directly to destination filesystem paths without proper normalization or containment checks.

Because of this, a user with write access to the source GCS bucket can create object names containing '..' segments or absolute path prefixes that cause the operators to write files outside the intended destination directories. This path traversal can lead to overwriting arbitrary files on the SFTP server or the worker host.

The vulnerability affects deployments where the GCS bucket is writable by less-trusted principals, such as partner uploads or ingest-only service accounts. The issue was fixed by adding validation to normalize and resolve destination paths, ensuring they remain within the configured base directory, and raising errors if a path escapes containment.

Useful 0 Not Useful 0

Impact Analysis

This vulnerability can allow an attacker with write access to the source GCS bucket to perform path traversal attacks by crafting malicious object names. As a result, the attacker could cause the Airflow DAG run to write files outside the intended destination directories.

• Overwrite arbitrary files on the SFTP server, potentially compromising data or system integrity.
• Overwrite files on the worker host, which could lead to unauthorized code execution, data corruption, or denial of service.

Overall, this could lead to unauthorized modification of files, system compromise, and disruption of workflows in affected Airflow deployments.

Useful 0 Not Useful 0

Detection Guidance

This vulnerability can be detected by monitoring for error messages raised by the affected operators when they encounter path traversal attempts. Specifically, the operators GCSToSFTPOperator and GCSTimeSpanFileTransformOperator now raise a ValueError if a GCS object name attempts to escape the configured destination directory after normalization.

To detect exploitation attempts, you can check Airflow logs for ValueError exceptions mentioning path traversal or suspicious object names containing '..' segments.

Additionally, you can audit the GCS bucket contents for object names containing '..' or absolute path segments that could be used to exploit this vulnerability.

• Use Airflow task logs to grep for ValueError related to path traversal, e.g., `grep -i 'ValueError' /path/to/airflow/logs/*`
• List GCS bucket objects with suspicious names using gsutil: `gsutil ls gs://your-bucket/** | grep '\.\.'`
• Check for unexpected file writes outside the intended destination directories on the SFTP server or worker host by monitoring filesystem changes or using tools like `find` to locate recently modified files outside expected paths.

Useful 0 Not Useful 0

Mitigation Strategies

The primary and recommended mitigation is to upgrade the apache-airflow-providers-google package to version 22.2.1 or later, which includes the fix that normalizes and validates destination paths to prevent path traversal.

Until the upgrade can be applied, restrict write access to the source GCS buckets to trusted principals only, preventing untrusted users from uploading objects with malicious names.

Review and audit your GCS buckets for any objects with suspicious names containing '..' segments and remove or rename them.

Monitor Airflow logs for ValueError exceptions related to path traversal attempts to detect exploitation attempts.

Useful 0 Not Useful 0

Compliance Impact

This vulnerability allows an attacker with write access to a Google Cloud Storage bucket to perform path traversal attacks, potentially overwriting arbitrary files on the SFTP server or worker host. Such unauthorized file overwrites could lead to data integrity and confidentiality issues.

While the provided information does not explicitly mention compliance with standards like GDPR or HIPAA, the ability to overwrite arbitrary files could result in unauthorized data access or modification, which may violate data protection and security requirements mandated by these regulations.

Therefore, organizations using affected versions of Apache Airflow's Google provider operators and ingesting data from less-trusted principals should upgrade to version 22.2.1 or later to mitigate risks that could impact compliance with such standards.

Useful 0 Not Useful 0