Signal is active with 1 mentions in latest observed window
Recommended action window: Monitor and triage in normal cycle
NVD description
Predictable bucket naming in Vertex AI Experiments in Google Cloud Vertex AI from version 1.21.0 up to (but not including) 1.133.0 on Google Cloud Platform allows an unauthenticated remote attacker to achieve cross-tenant remote code execution, model theft, and poisoning via pre-creating predictably named Cloud Storage buckets (Bucket Squatting). This vulnerability was patched and no customer action is needed.
Priority
LOW
Exploitation
NONE
PoC
NONE
Patch
AVAILABLE
Momentum
STABLE
| Date | Total | Labels |
|---|
| 2026-02-20 | 1 | Disclosure1 |
| 2026-03-01 | 1 | Disclosure1 |
| 2026-04-20 | 1 | Disclosure1 |
| 2026-06-16 | 1 | Patch1 |
| 2026-06-18 | 1 | Disclosure1 |
| 2026-06-23 | 1 | False Positive1 |
| 2026-07-10 | 1 | General1 |
... la interpretación del Reglamento (UE) 2023/1115 es espuria y, por lo tanto, solicitamos que se retire del curso legal la Resolución CVE-2026-2473 del Gobierno de Cantabria del PP de Buruaga.
Post summary
The text debunks the validity of the Resolution CVE‑2026‑2473, claiming it is spurious, with no other actionable information about the vulnerability.
I found this squatting pattern recurring across three major GCP services: • VertexSquat (CVE-2026-2473): Full RCE in Vertex AI. • GeminiSquat (CVE-2026-1727): Gemini Enterprise. • MountSquat: Takeover of Cloud Run mount volumes.
Post summary
The post identifies a pattern of vacated security reviews across Google Cloud services, highlighting CVE-2026-2473 as a full RCE in Vertex AI, while noting related CVEs without providing exploit, mitigation, or PoC details.
TRC analysis shows attackers exploiting predictable bucket naming patterns to silently redirect organizational data streams to attacker-controlled storage. CVE-2026-2473 and CVE-2026-1727 demonstrate how misconfigured cloud storage enables privilege escalation and lateral movement across cloud resources. Runtime segmentation could help contain post-compromise activity. #CloudSecurity 🔗 Full TRC analysis: https://aviatrix.ai/threat-research-center/threatsday-cloud-bucket-hijacking-2026
Post summary
The analysis details how predictable bucket naming patterns can lead to misconfigured cloud storage, referencing CVE-2026-2473 and CVE-2026-1727, but offers no PoC, exploit code, patches, or evidence of active exploitation.
Unit 42 disclosed CVE-2026-2473 in Google Cloud Vertex AI SDK — bucket squatting + Pickle deserialization enables cross-tenant RCE with zero access to the victim’s GCP project. Patched in google-cloud-aiplatform v1.148.0. https://dailysecurityreview.com/resources/cve-2026-2473-vertex-ai-sdk-pickle-attack-enables-cross-tenant-rce/ #CyberSecu https://t.co/fBwmz33Pux
Post summary
Unit 42 disclosed CVE‑2026‑2473, a bucket squatting and Pickle deserialization flaw in Google Cloud Vertex AI SDK that enables cross‑tenant RCE without project access, and the issue is patched in google‑cloud‑aiplatform v1.148.0.
🚨 UPDATE — Vertex AI bucket squatting: This isn't a one-off. Focal Security separately disclosed THREE Google Cloud bucket squatting flaws this year — GeminiSquat (CVE-2026-1727) in Gemini Enterprise, MountSquat in Cloud Run, and VertexSquat (CVE-2026-2473) in Vertex AI Experiments. All three: no credentials needed, cross-tenant code execution. No CVE has been assigned to the new 'Pickle in the Middle' flaw yet. But the pattern is clear: Google Cloud's predictable bucket naming convention is a systemic architectural problem, not a one-time bug. Check your google-cloud-aiplatform version in notebooks, CI jobs, and training pipelines — not just production. Update to v1.148.0 now. 👇 https://thehackernews.com/2026/06/google-vertex-ai-sdk-flaw-let-attackers.html
Post summary
Three Google Cloud bucket squatting vulnerabilities enabled cross‑tenant code execution without credentials; users are urged to update to v1.148.0.
Vertex AI Security Alert CVE-2026-2473: Predictable bucket naming in Vertex AI Experiments could allow cross-tenant RCE & model theft (Bucket Squatting). Official Google Cloud Bulletin: NVD Entry: https://nvd.nist.gov/vuln/detail/CVE-2026-2473 #VertexAI #GoogleCloud #CloudSecurity #CVE #DevSecOps
Post summary
Google Cloud has issued a security alert for CVE-2026-2473, noting that predictable bucket naming in Vertex AI Experiments could lead to cross‑tenant remote code execution and model theft.
CVE-2026-2473 Predictable bucket naming in Vertex AI Experiments in Google Cloud Vertex AI from version 1.21.0 up to (but not including) 1.133.0 on Google Cloud Platform allows an un… https://www.cve.org/CVERecord?id=CVE-2026-2473
Post summary
The text provides a brief disclosure of a predictable bucket naming issue in Vertex AI Experiments across specific versions, highlighting a potential security risk.