Remote code execution via unverified staging bucket
Published Sep 15, 2026 · Updated Sep 15, 2026
Bucket squatting in Google Cloud Agent Platform SDK for Python before 1.165.1 allows remote attackers to execute code in model serving. The stage_local_data_in_gcs() path in gcs_utils.py derives a predictable default bucket name and accepts an existing bucket without verifying ownership, letting an attacker replace uploaded model artifacts. Exploitation requires a new regional default bucket, no explicit staging_bucket, knowledge of the project ID, and winning the replacement race before the service agent loads the poisoned model; execution can expose its tenant-project token.
Summary
What happened
Bucket squatting in Google Cloud Agent Platform SDK for Python before 1.165.1 allows remote attackers to execute code in model serving. The stage_local_data_in_gcs() path in gcs_utils.py derives a predictable default bucket name and accepts an existing bucket without verifying ownership, letting an attacker replace uploaded model artifacts. Exploitation requires a new regional default bucket, no explicit staging_bucket, knowledge of the project ID, and winning the replacement race before the service agent loads the poisoned model; execution can expose its tenant-project token.
The record
- CVE
- CVE-2026-19407
- Published
- Sep 15, 2026
- Updated
- Sep 15, 2026
- Vendor
- Google Cloud
- Product
- Agent Platform SDK for Python
- Classifications
- T1195.002
- Attack vector
- network
- Privileges
- unauthenticated
Timeline
How it unfolded
- Sep 15, 2026CVE publishedPublication date reported by the CVE source.
- Sep 15, 2026Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=1.16.0 <1.165.1
What conditions does exploitation require?
What is affected?
Published CVSS scores
No CVSS assessment is available in this record.
CVSS describes severity. EPSS estimates exploitation probability.
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Weakness, pattern, technique
Public exploit references
- Pickle in the Middle proof of conceptproof of concept · demonstrated
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo