CVE-2026-23529
Prior to 2.11.0, there is an arbitrary file read in Google BigQuery Sink connector.
Does this matter?
High impact if exploited, but EPSS currently rates exploitation as unlikely (0.39%). Schedule it in the normal patch cycle and watch for a rise in EPSS or a public exploit.
Description
Kafka Connect BigQuery Connector is an implementation of a sink connector from Apache Kafka to Google BigQuery. Prior to 2.11.0, there is an arbitrary file read in Google BigQuery Sink connector. Aiven's Google BigQuery Kafka Connect Sink connector requires Google Cloud credential configurations for authentication to BigQuery services. During connector configuration, users can supply credential JSON files that are processed by Google authentication libraries. The service fails to validate externally-sourced credential configurations before passing them to the authentication libraries. An attacker can exploit this by providing a malicious credential configuration containing crafted credential_source.file paths or credential_source.url endpoints, resulting in arbitrary file reads or SSRF attacks.
- CVSS 3.1
- 7.7 HIGHCVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N
- EPSS
- 0.39% probability · 33th percentile
- CISA KEV
- Not listed
- Weakness
- CWE-73, CWE-918
- Source
- security-advisories@github.com
References
- https://docs.cloud.google.com/support/bulletins#gcp-2025-005
- https://github.com/Aiven-Open/bigquery-connector-for-apache-kafka/commit/20ea3921c6fe72d605a033c1943b20f49eaba981
- https://github.com/Aiven-Open/bigquery-connector-for-apache-kafka/releases/tag/v2.11.0
- https://github.com/Aiven-Open/bigquery-connector-for-apache-kafka/security/advisories/GHSA-3mg8-2g53-5gj4
Source: NVD record, EPSS from FIRST.org, KEV from CISA. Refreshed daily.