CVE-2026-19485 PUBLISHED

Bucket Squatting in Vertex AI Search for Commerce

Assigner: GoogleCloud
Reserved: 10.08.2026 Published: 26.08.2026 Updated: 26.08.2026

A Predictable Resource Name vulnerability in BigQuery Import Staging in Google Cloud Vertex AI Search for Commerce versions prior to 2026-04-27 on Google Cloud Platform allows an attacker knowing the victim's project number to obtain read/write access to staged data and error logs using predictable bucket names.

This vulnerability was patched and no customer action is needed.

Metrics

CVSS 4.0

CVSS Vector: CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N/U:Clear
CVSS Score: 9.3

Exploitability Metrics		Vulnerable System Impact Metrics		Subsequent System Impact Metrics
Attack Vector	Network	Confidentiality	High	Confidentiality	High
Attack Complexity	Low	Integrity	High	Integrity	High
Attack Requirements	None	Availability	None	Availability	None
Privileges Required	Low
User Interaction	None

CVSS 4.0

Product Status

Vendor	Google Cloud
Product	Vertex AI Search for Commerce
Versions	Default: unaffected affected from 0 to 2026-04-27 (excl.)

Credits

Omer Amiad reporter

References

https://unit42.paloaltonetworks.com/hijacking-vertex-ai-model/

Problem Types

CWE-330 Use of Insufficiently Random Values CWE

Impacts

CAPEC-616 Establish Rogue Location