Severity by source
AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
Local vector and UI:R confirmed by description; PR:N reflects no prior privileges needed; full C/I/A High consistent with privilege escalation to arbitrary command execution.
Primary rating from Vendor (microsoft).
CVSS VectorVendor: microsoft
Lifecycle Timeline
3DescriptionCVE.org
Improper neutralization of special elements used in an os command ('os command injection') in GitHub Copilot and Visual Studio Code allows an unauthorized attacker to elevate privileges locally.
AnalysisAI
OS command injection in GitHub Copilot and Visual Studio Code enables a local unauthorized attacker to escalate privileges on the affected developer workstation, achieving full High impact across confidentiality, integrity, and availability. The CWE-78 root cause indicates that attacker-influenced input containing shell metacharacters is improperly incorporated into OS-level command execution within one or both products, triggering upon a user action such as opening a malicious file or repository. No public exploit code has been identified and this vulnerability has not been confirmed actively exploited by CISA KEV at time of analysis; however, the local privilege escalation path on widely deployed developer tooling represents a meaningful risk in supply chain and social engineering scenarios.
Technical ContextAI
CWE-78 (Improper Neutralization of Special Elements used in an OS Command) describes vulnerabilities where attacker-controlled strings - containing characters such as semicolons, pipes, backticks, or newlines - are concatenated into OS shell commands without adequate sanitization or quoting. In the context of GitHub Copilot and Visual Studio Code, the injection surface likely resides in how one or both products process repository content, file paths, workspace configuration, extension inputs, or AI-analyzed code snippets that are subsequently passed to a subprocess or shell API. The CVSS local attack vector (AV:L) and user interaction requirement (UI:R) are consistent with a delivery path involving malicious workspace or file content processed during normal developer activity. No CPE strings were provided in the available intelligence, so exact affected product versions and build numbers must be derived directly from the MSRC advisory.
RemediationAI
Consult the Microsoft Security Response Center advisory at https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-70335 for officially confirmed patched versions of Visual Studio Code and the GitHub Copilot extension; specific fix version numbers were not present in the available intelligence data and should not be inferred. As an interim compensating control, apply VS Code workspace trust settings to restrict automatic execution of tasks, terminal profiles, and extension features when opening untrusted folders or repositories - this directly targets the UI:R delivery vector by requiring explicit user approval before potentially dangerous workspace actions execute. Additionally, avoid opening files or workspaces from unknown or untrusted sources, particularly those shared externally. Note that workspace trust alone does not patch the underlying command injection flaw and should be treated as a stopgap until vendor-released updates are applied.
More in Visual Studio Code
View allVisual Studio Code Remote Code Execution Vulnerability. Rated high severity (CVSS 7.8), this vulnerability is no authent
Visual Studio Code Remote Code Execution Vulnerability. Rated high severity (CVSS 8.8), this vulnerability is remotely e
Security-feature bypass in Microsoft Visual Studio Code before 1.136.2 lets a remote, unauthenticated attacker defeat a
Privilege escalation in Microsoft Visual Studio Code allows remote unauthenticated attackers to elevate privileges over
Security-feature bypass in Microsoft Visual Studio Code lets a remote attacker deliver functionality from an untrusted c
Remote code execution in Visual Studio Code (versions 1.0.0 through 1.132.0) is enabled by a missing authorization check
Remote code execution in Microsoft Visual Studio Code allows a network-based attacker to inject and execute arbitrary co
Authorization bypass in Visual Studio Code versions 1.0.0 through 1.136.1 allows an unauthenticated network attacker to
OS command injection in Visual Studio Code exposes victims to remote code execution when they interact with attacker-con
Security-feature bypass in GitHub Copilot and Visual Studio (affecting Visual Studio Code builds before 1.119.1) lets a
Privilege elevation in Microsoft Visual Studio Code (all releases from 1.0.0 up to but not including 1.119.1) stems from
Missing authentication for critical function in Visual Studio Code extension for Arduino allows an unauthenticated attac
Same weakness CWE-78 – OS Command Injection
View allSame technique Command Injection
View allShare
External POC / Exploit Code
Leaving vuln.today
EUVD-2026-56588
GHSA-hj52-q5wj-vpwg