CVE-2026-92762: missing authz lets low-priv users run

CISO Take

Pelican Panel versions before 1.0.0-beta35 enforce startup write permissions only through disabled form controls in the UI, so any user with read-only startup access can forge Livewire state updates and change startup commands, Docker images and variables to execute arbitrary commands in the container. The CVSS score is 8.8 (network, low complexity, low privileges, no user interaction), but there is no public exploit, no Nuclei template, no CISA KEV listing and a low EPSS of 0.00654, and CISA SSVC rates it ATTEND, so this is a prioritized patch rather than an emergency. The risk is concentrated in multi-tenant game, AI and ML hosting setups where low-trust tenants hold panel accounts, because a read-only role becomes code execution on the workload. Upgrade to 1.0.0-beta35 or later, and until then review who holds startup.read and audit recent changes to startup commands and images.

Sources:NVDEPSSGitHub AdvisoryATLASvulncheck.com

What is the risk?

High severity (CVSS 8.8), but exploitation likelihood is currently low: EPSS is 0.00654, with no public exploit, no scanner template and no KEV listing. Exploitation is trivial for any authenticated user with startup.read, since the flaw is a missing server-side authorization check (CWE-862) and the attacker only needs to craft Livewire requests. Exposure depends on whether the panel is internet-facing and how many low-trust accounts exist. AI relevance is indirect: Pelican is a server management panel, categorized here as ml_ui, that can host AI/ML workloads and inference containers.

How does the attack unfold?

Initial Access

Attacker authenticates to Pelican Panel with a low-privilege account holding startup.read.

AML.T0012

Exploitation

Attacker forges Livewire state updates that trigger afterStateUpdated callbacks, bypassing the missing server-side authorization check.

AML.T0049

Execution

Modified startup commands, Docker image or variables run attacker-controlled commands in the container.

AML.T0050

Impact

Attacker steals secrets, tampers with hosted AI workloads and abuses compute resources.

Initial Access

Attacker authenticates to Pelican Panel with a low-privilege account holding startup.read.

AML.T0012

Exploitation

Attacker forges Livewire state updates that trigger afterStateUpdated callbacks, bypassing the missing server-side authorization check.

AML.T0049

Execution

Modified startup commands, Docker image or variables run attacker-controlled commands in the container.

AML.T0050

Impact

Attacker steals secrets, tampers with hosted AI workloads and abuses compute resources.

How severe is it?

CVSS 3.1

8.8 / 10

EPSS

0.7%

chance of exploitation in 30 days

Higher than 51% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Trivial

Exploitation Confidence

medium

○CISA SSVC: Public PoC

○Public PoC indexed (trickest/cve)

Composite signal derived from CISA KEV, VulnCheck KEV, CISA SSVC, EPSS, Metasploit, Exploit-DB, trickest/cve, Nuclei templates, and inthewild.io exploitation reports.

What is the attack surface?

AVNetwork

ACLow

PRLow

UINone

SUnchanged

CHigh

IHigh

AHigh

What should I do?

1 step

1) Upgrade Pelican Panel to 1.0.0-beta35 or later. 2) Until patched, restrict the panel to trusted networks or a VPN and remove startup.read from untrusted roles. 3) Audit server startup commands, Docker images and variables for unexpected changes, and review panel and Livewire request logs for startup state updates from users without write permission. 4) Rotate secrets and tokens exposed as environment variables in any container where unauthorized changes are found. 5) Harden containers with least privilege, no privileged mode, restricted networking and read-only mounts to limit blast radius.

What does CISA's SSVC say?

DecisionAttend

Exploitationpoc

AutomatableNo

Technical Impacttotal

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

Auth BypassCode ExecutionFrameworkInferenceAML.T0012- Valid Accounts AML.T0049- Exploit Public-Facing Application AML.T0050- Command and Scripting Interpreter

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Art. 15-Accuracy, robustness and cybersecurity

ISO 42001

A.6.2.6-AI system operation and monitoring

NIST AI RMF

MANAGE 2.3-Mechanisms to respond to and recover from unknown risks

Frequently Asked Questions

What is CVE-2026-92762?

Pelican Panel versions before 1.0.0-beta35 enforce startup write permissions only through disabled form controls in the UI, so any user with read-only startup access can forge Livewire state updates and change startup commands, Docker images and variables to execute arbitrary commands in the container. The CVSS score is 8.8 (network, low complexity, low privileges, no user interaction), but there is no public exploit, no Nuclei template, no CISA KEV listing and a low EPSS of 0.00654, and CISA SSVC rates it ATTEND, so this is a prioritized patch rather than an emergency. The risk is concentrated in multi-tenant game, AI and ML hosting setups where low-trust tenants hold panel accounts, because a read-only role becomes code execution on the workload. Upgrade to 1.0.0-beta35 or later, and until then review who holds startup.read and audit recent changes to startup commands and images.

Is CVE-2026-92762 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2026-92762, increasing the risk of exploitation.

How to fix CVE-2026-92762?

1) Upgrade Pelican Panel to 1.0.0-beta35 or later. 2) Until patched, restrict the panel to trusted networks or a VPN and remove startup.read from untrusted roles. 3) Audit server startup commands, Docker images and variables for unexpected changes, and review panel and Livewire request logs for startup state updates from users without write permission. 4) Rotate secrets and tokens exposed as environment variables in any container where unauthorized changes are found. 5) Harden containers with least privilege, no privileged mode, restricted networking and read-only mounts to limit blast radius.

What systems are affected by CVE-2026-92762?

This vulnerability affects the following AI/ML architecture patterns: Model serving, Containerized AI/ML hosting panels, Multi-tenant compute management.

What is the CVSS score for CVE-2026-92762?

CVE-2026-92762 has a CVSS v3.1 base score of 8.8 (HIGH). The EPSS exploitation probability is 0.65%.

What is the AI security impact?

Affected AI Architectures

Model servingContainerized AI/ML hosting panelsMulti-tenant compute management

MITRE ATLAS Techniques

AML.T0012Valid Accounts

AML.T0049Exploit Public-Facing Application

AML.T0050Command and Scripting Interpreter

Compliance Controls Affected

EU AI Act: Art. 15

ISO 42001: A.6.2.6

NIST AI RMF: MANAGE 2.3

What are the technical details?

Original Advisory

Pelican Panel versions before 1.0.0-beta35 enforce startup write permissions only through disabled form controls rather than server-side authorization checks. Attackers with startup.read permission can craft Livewire state updates to invoke afterStateUpdated callbacks and modify startup commands, docker images, and variables to execute arbitrary commands in the container.

Exploitation Scenario

An attacker holds a low-privilege sub-user account with startup.read on a Pelican-managed server that runs an LLM inference service. They open the Startup page, where the controls are disabled in the UI, and replay a crafted Livewire update that triggers the afterStateUpdated callbacks. They change the startup command to launch a reverse shell, or swap the Docker image for a malicious one, and restart the server. The attacker then reads API keys and model weights from the container and uses the compute for their own purposes.

Weaknesses (CWE)

CWE-862Missing AuthorizationPrimary CWE-862Missing Authorization

CWE-862 — Missing Authorization: The product does not perform an authorization check when an actor attempts to access a resource or perform an action.

[Architecture and Design] Divide the product into anonymous, normal, privileged, and administrative areas. Reduce the attack surface by carefully mapping roles with data and functionality. Use role-based access control (RBAC) [REF-229] to enforce the roles at the appropriate boundaries. Note that this approach may not protect against horizontal authorization, i.e., it will not protect a user from attacking others with the same role.
[Architecture and Design] Ensure that access control checks are performed related to the business logic. These checks may be different than the access control checks that are applied to more generic resources such as files, connections, processes, memory, and database records. For example, a database may restrict access for medical records to a specific database user, but each record might only be intended to be accessible to the patient and the patient's doctor [REF-7].

Source: MITRE CWE corpus.