EUVD-2026-48230
GHSA-662q-mj5f-wv34
Severity by source
AV:L/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Code runs as the developer with full impact via local delivery, but a victim must open the attacker's project, so AV:L and UI:R; no privileges needed (PR:N).
Primary rating from Vendor (jetbrains).
CVSS VectorVendor: jetbrains
Lifecycle Timeline
3DescriptionCVE.org
In JetBrains WebStorm before 2026.2 arbitrary code execution was possible before granting project trust via project-local package-manager tooling
AnalysisAI
Arbitrary code execution in JetBrains WebStorm before 2026.2 lets a malicious project run code on a developer's machine before the IDE's project-trust prompt is ever answered, by abusing project-local package-manager tooling that the IDE invokes automatically. Because execution happens prior to the trust gate, the safety boundary meant to protect users who merely open an untrusted repository is bypassed. There is no public exploit identified at time of analysis and it is not listed in CISA KEV, but the RCE tag and full C/I/A impact make this a meaningful developer-workstation risk.
Technical ContextAI
WebStorm is JetBrains' JavaScript/TypeScript IDE, built on the IntelliJ platform, which integrates tightly with Node.js package managers (npm, Yarn, pnpm) and their configuration files. The flaw is classified as CWE-829 (Inclusion of Functionality from Untrusted Control Sphere): when a project is opened, the IDE resolves and invokes package-manager tooling defined by files inside the project directory itself (for example package.json scripts, lockfiles, or a project-local package-manager binary/path). That project-controlled tooling is trusted and executed before the user grants project trust, so attacker-supplied configuration in a cloned or downloaded repository is pulled into the execution sphere of the IDE process running as the developer's user account.
RemediationAI
Vendor-released patch: upgrade WebStorm to 2026.2 or later, which closes the pre-trust execution path; obtain it from JetBrains Toolbox or via the advisory at https://www.jetbrains.com/privacy-security/issues-fixed/. Until updated, the actionable compensating control is to avoid opening untrusted or third-party projects in WebStorm and, where the workflow allows, inspect a repository's package.json scripts, lockfiles, and any project-local package-manager configuration in a plain text editor or sandbox/VM before opening it in the IDE (trade-off: slower onboarding of external code and no protection if the review misses a malicious script). Running the IDE inside a disposable container or VM for untrusted-code triage limits blast radius but adds workflow friction; disabling automatic package-manager resolution features where configurable reduces exposure at the cost of losing dependency-tooling convenience.
In JetBrains WebStorm before 2021.1, code execution without user confirmation was possible for untrusted projects. Rated
Arbitrary code execution in JetBrains WebStorm before 2026.2 allows a malicious project to run code through the configur
Arbitrary code execution in JetBrains WebStorm before 2026.2 allows a malicious project to run code via project-local li
Arbitrary code execution in JetBrains WebStorm before 2026.2 allows a malicious project to run attacker-controlled code
In JetBrains WebStorm before 2024.3 code execution in Untrusted Project mode was possible via type definitions installer
GitHub access token could be exposed to third-party sites in JetBrains IDEs after version 2023.1 and less than: IntelliJ
In JetBrains WebStorm before 2021.1, HTTP requests were used instead of HTTPS. Rated high severity (CVSS 7.5), this vuln
Share
External POC / Exploit Code
Leaving vuln.today