Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessObtain local low-privileged account
-
DeliveryIdentify NemoClaw Telegram bridge process
-
ExploitCraft shell-metacharacter payload
-
ExecutionSubmit payload via bridge input path
-
PersistTrigger unsanitized OS command execution
-
ImpactAchieve code execution or privilege escalation
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A05:2025 - Injection
ASVS 5.0
V1 - Encoding and Sanitization OS command sanitization, avoid shell exec
MITRE ATT&CK
2 techniques across 1 tactic: Execution The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | The Telegram bridge component of NemoClaw must be enabled and running; deployments that do not use the Telegram integration are not exposed through this specific vector. A local user account with at least low-level system privileges (PR:L) is required - fully unauthenticated remote exploitation is not supported by the CVSS vector (AV:L). No user interaction from a second party is needed (UI:N), so the attacker can trigger the condition unilaterally once they have local access and can supply crafted input to the bridge. Exploitation is limited to Linux; Windows or other platform deployments are not covered by this CVE. The wildcard CPE implies no specific configuration flag beyond bridge enablement is documented as a prerequisite. |
| Impact | The CVSS 3.1 vector CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H scores 7.8, reflecting local network access, low complexity, and low privilege requirements with high impact across all three pillars. The local attack vector (AV:L) limits opportunistic internet-facing exploitation, but any multi-tenant or shared Linux system running NemoClaw with the Telegram bridge enabled becomes a lateral movement target once an attacker achieves any foothold on the host. The unchanged scope (S:U) suggests impact is bounded to the vulnerable system itself, not cascading to adjacent systems via the vulnerability mechanism alone. No EPSS percentile data was provided in the source intelligence, and the CVE is not listed in CISA KEV, so no confirmed active exploitation signal exists. The absence of a public POC further tempers urgency relative to a KEV-confirmed issue. Nevertheless, AC:L and PR:L together mean exploitation is straightforward for any low-privileged local user, making this a meaningful internal privilege escalation risk in shared or containerized AI workload environments - even a mid-sized deployment team running shared GPU nodes should treat this as urgent pending patch availability. |
| Exploit ScenarioDerived | A low-privileged user on a shared Linux GPU node running NemoClaw with the Telegram bridge enabled crafts a malicious string - such as a specially formatted Telegram message or a manipulated configuration value - that is passed unsanitized to a shell command within the bridge component. The injected shell metacharacters cause the application to execute attacker-controlled commands under the NemoClaw process identity, potentially granting access to higher-privileged files or enabling persistence mechanisms. No public proof-of-concept code has been identified, but the low complexity rating indicates the technique does not require specialized tooling. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Aug 25, 2026
OS command injection in NVIDIA NemoClaw's Telegram bridge component on Linux enables a local attacker with low-privileged access to execute arbitrary operating system commands. Exploitation yields full confidentiality, integrity, and availability compromise on the affected host, with potential for privilege escalation beyond the invoking user. No active exploitation is confirmed in CISA KEV and no public exploit code has been identified at time of analysis, but the low attack complexity (AC:L) and minimal privilege requirement (PR:L) lower the practical bar for exploitation once local access is obtained.
RemediationAI
Consult the NVIDIA Product Security advisory at https://github.com/NVIDIA/product-security/tree/main/2026/5872 for the authoritative patch release; no specific fixed version number was confirmed in available intelligence at time of analysis. As an immediate compensating control, disable or remove the Telegram bridge component if it is not operationally required, which eliminates the vulnerable code path entirely. If the bridge must remain active, restrict local user access to the NemoClaw process to the minimum necessary set of accounts, and apply OS-level mandatory access controls (AppArmor or SELinux policies) to constrain the process's ability to spawn child shells. Auditing and alerting on unexpected subprocess invocations from the NemoClaw process can provide detection coverage until a vendor patch is applied. Note that access-restriction controls reduce attack surface but do not remediate the underlying injection flaw.
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
CWE-78 (Improper Neutralization of Special Elements used in an OS Command) describes a class of injection flaws in which attacker-controlled input is concatenated into, or otherwise influences, a shell command without sufficient escaping or allowlisting. In the context of NVIDIA NemoClaw's Telegram bridge, the affected code path plausibly passes message content, configuration parameters, or user-supplied identifiers into a subprocess or shell invocation unsanitized. The CPE string cpe:2.3:a:nvidia:nemoclaw:*:*:*:*:*:*:*:* covers all known versions of the software with no version floor or ceiling specified, indicating the full release history is considered affected. NemoClaw is an NVIDIA AI application toolkit for Linux environments, and the Telegram bridge component provides messaging integration, a surface area that commonly handles externally-sourced string input.
Know when Nemoclaw 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
AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
Local vector confirmed by component architecture; low privilege sufficient to reach bridge input; no user interaction needed; full CIA impact from arbitrary command execution.
Primary rating from Vendor (nvidia).
CVSS VectorVendor: nvidia
DescriptionCVE.org
NVIDIA NemoClaw for Linux contains a vulnerability in the Telegram bridge component, where an attacker could cause an OS command injection. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, information disclosure, and data tampering.
Articles & Coverage 1
Executive ViewAI
NVIDIA NemoClaw's Telegram bridge component on Linux contains an OS command injection vulnerability (CVE-2026-65096, CVSS 7.8) enabling low-privileged local users to execute arbitrary system commands with full host compromise potential including data breach and privilege escalation. No public exploit identified at time of analysis, but no vendor patch is available, leaving affected organizations exposed to insider threats and compromised low-level accounts.
No vendor-released patch identified at time of analysis. Interim mitigations: disable Telegram bridge if operationally feasible; restrict local user access through OS-level account creation and privilege controls; implement network segmentation to contain affected systems; deploy command execution and file integrity monitoring on NemoClaw hosts; audit and minimize Telegram API permissions; maintain detailed audit logs of all bridge communications for forensic capability.
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
OS command injection occurs when an application passes unsanitized user input directly into system shell commands.
Information disclosure occurs when an application unintentionally exposes sensitive data that aids attackers in reconnaissance or directly compromises security.
Remote Code Execution represents the critical moment when an attacker successfully runs arbitrary code on a target system without physical access.
Weak authentication in NVIDIA NemoClaw for Linux (versions 0 through 0.0.4) lets remote attackers bypass the authenticat
Improper certificate validation in NVIDIA NemoClaw for Linux (versions up to and including 0.0.1) allows network attacke
Untrusted code execution in NVIDIA NemoClaw for Linux (versions up to and including 0.0.21) stems from the installer exe
Unverified code download in NVIDIA NemoClaw installation scripts for Linux (versions 0.0.0 through 0.0.21) allows a netw
Remote unauthenticated attackers can exfiltrate sensitive host environment variables from NVIDIA NeMoClaw by injecting m
Unauthenticated access to the NVIDIA NemoClaw inference server on Linux exposes sensitive model inference data and enabl
OS command injection in NVIDIA NemoClaw for Linux allows a local low-privileged attacker to execute arbitrary operating
OS command injection in NVIDIA NemoClaw's NIM management component on Linux enables a local authenticated attacker to ex
OS command injection in NVIDIA NemoClaw's command-line interface on Linux enables a local low-privileged attacker to exe
Code injection in NVIDIA NemoClaw's migration command (Linux, versions up to 0.0.17) allows a local attacker with low-pr
Server-side request forgery in NVIDIA NemoClaw's validateEndpointUrl() function allows local attackers with user interac
Credential exposure in NVIDIA NemoClaw allows a local low-privileged user, upon triggering user interaction, to access i
EUVD-2026-66093
GHSA-4qrf-4qfq-jhwq
NVD
CVE.org
EPSS
Vulnrichment
Exploit-DB