Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessLure victim with active Cloud Shell session to malicious page
-
DeliveryPage sends cross-origin WebSocket request
-
ExploitOrigin validation bypassed, session hijacked
-
ExecutionExfiltrate OAuth access token
-
ImpactImpersonate user against GCP resources
Vulnerability Assessment
| Weakness |
CWE
ASVS 5.0
V15 - Secure Coding and Architecture Hardware safety mechanism
|
| Exploitation Conditions | Exploitation requires two specific preconditions: (1) the victim must have an active, authenticated Google Cloud Shell session open, and (2) the victim must be induced to load an attacker-controlled web page that issues a cross-origin WebSocket request (the UI:P passive-interaction factor in the CVSS vector). The underlying technical condition is the Cloud Shell WebSocket endpoint's failure to validate the request Origin (CWE-1385). The attacker needs no credentials or prior access (PR:N). Limiting factors: the required user interaction and the need for a concurrently active session narrow the window, and because Google deployed a server-side fix on 2026-07-22 the vulnerable code path no longer exists in the hosted service - there is no on-premises or customer-patchable component. |
| Impact | This is a genuinely high-risk design flaw, but with important real-world caveats. The CVSS 4.0 base score of 8.7 is driven by a network attack vector (AV:N), low complexity (AC:L), no attacker privileges (PR:N), and high impact to confidentiality, integrity, and availability - justified because a stolen OAuth access token effectively grants the attacker the victim's Cloud Shell identity. The single factor that meaningfully constrains exploitation is the passive user-interaction requirement (UI:P): the victim must have an active Cloud Shell session and be lured to an attacker-controlled web page that issues the cross-origin WebSocket request. No EPSS score, CISA KEV listing, or public POC is present in the provided data, so by the precision rules this is classified as 'no public exploit identified at time of analysis' - the elevated score reflects latent potential and the sensitivity of the OAuth token, not observed exploitation. Crucially, the vendor states the flaw was fixed server-side on 2026-07-22 with no customer action required, so for the hosted Cloud Shell service the practical residual risk today is effectively zero; the analytical risk applies to the pre-patch window. No signals conflict, though note the CVSS 4.0 subsequent-system metrics (SC/SI/SA:L) imply the token enables impact on resources beyond Cloud Shell itself. |
| Exploit ScenarioDerived | A victim with an active Google Cloud Shell session is tricked into visiting a malicious web page (satisfying the UI:P requirement). The page opens a cross-origin WebSocket connection to the Cloud Shell backend, which fails to validate the request origin (CWE-1385), hijacking the session and exfiltrating the user's OAuth access token to the attacker. No public exploit code is identified in the provided data, and Google reports the issue was fixed server-side on 2026-07-22. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Oct 20, 2026
Cross-site WebSocket hijacking in Google Cloud Shell prior to 2026-07-22 lets an unauthenticated remote attacker hijack an active, authenticated Cloud Shell session and exfiltrate the user's OAuth access token. Exploitation requires the victim to have a live Cloud Shell session and to load an attacker-controlled page that issues a cross-origin WebSocket request, which is the user-interaction precondition reflected in the CVSS vector. Google remediated the hosted service server-side on 2026-07-22, and no public exploit was identified at time of analysis.
RemediationAI
Primary remediation is vendor-side: Google patched the hosted Cloud Shell service on 2026-07-22, and no customer action is required because there is no customer-patchable component; confirm details in the release notes at https://docs.cloud.google.com/shell/docs/release-notes#July_28_2026. For assurance around sessions that predate the fix, close all active Cloud Shell sessions and reauthenticate so new sessions use the fixed endpoint, and revoke or reissue OAuth tokens if suspicious activity is suspected; this disrupts active work but limits any residual session hijack window. As a compensating control while unpatched exposure is suspected, avoid loading untrusted web pages in a browser profile that has an active Cloud Shell session and close idle Cloud Shell sessions; this reduces the cross-origin WebSocket attack surface but does not remediate the server-side flaw and can interfere with normal workflow. No customer-installable patch version beyond the 2026-07-22 service update is provided.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
Google Cloud Shell is a browser-accessed, hosted terminal service; its client communicates with Google infrastructure over WebSocket connections, and the session carries the user's OAuth access token. The root cause is CWE-1385, missing Origin validation in WebSockets: the Cloud Shell WebSocket endpoint did not adequately verify the Origin of cross-origin requests, allowing a malicious web page to open a WebSocket to the service in the context of the victim's active session and read or relay session data. Because WebSockets are not governed by the same-origin policy in the same way as ordinary XMLHttpRequest/fetch, Origin checking is the key control preventing cross-site WebSocket hijacking. The affected component is the hosted Google Cloud Shell service, tracked by CPE cpe:2.3:a:google_cloud:cloud_shell:*:*:*:*:*:*:*:*, so there is no customer-managed package or on-premises code path to patch. The CVSS 4.0 vector has PR:N, meaning no attacker credentials are required, and UI:P, reflecting passive victim interaction; the independent assessment maps this to CVSS 3.1 AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H, with the main limiting factors being the need for a concurrently active Cloud Shell session and user interaction.
Know when Cloud Shell gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:L/SI:L/SA:L/U:Clear
Unauthenticated remote request (PR:N, AV:N) needs victim to visit a malicious page (UI:R); stolen OAuth token crosses into the user's GCP resources (S:C) with full C/I/A impact.
Primary rating from Vendor (GoogleCloud).
CVSS VectorVendor: GoogleCloud
Lifecycle Timeline
2DescriptionCVE.org
A Cross-Site WebSocket Hijacking vulnerability in Cloud Shell prior to 2026-07-22 on Google Cloud Platform allows an unauthenticated attacker to hijack an active Cloud Shell session and exfiltrate the user's OAuth access token using a cross-origin WebSocket request.
This vulnerability was patched on 22 July 2026, and no customer action is needed.
Executive ViewAI
Google Cloud Shell sessions could be hijacked through a cross-site WebSocket flaw, enabling theft of users' OAuth access tokens and potential unauthorized access to cloud resources. Google applied a server-side remediation on 2026-07-22 and no public exploit was identified at time of analysis, but leadership should care because compromised Cloud Shell credentials can expose sensitive cloud environments and administrative capabilities. Engineering and cloud operations teams should verify remediation and assess whether any active sessions or tokens require revocation.
No customer-applied patch; Google remediated the hosted Cloud Shell service server-side on 2026-07-22. Interim mitigations include restricting Cloud Shell use to trusted pages, shortening OAuth token lifetimes, revoking and reissuing tokens for active sessions, requiring reauthentication, and monitoring for anomalous token use.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Information disclosure occurs when an application unintentionally exposes sensitive data that aids attackers in reconnaissance or directly compromises security.
EUVD-2026-104858
GHSA-342c-3cqc-4gph
NVD
CVE.org
EPSS
Vulnrichment
Exploit-DB