Description

In JetBrains IntelliJ IDEA before 2026.3, 2026.2.4 command execution via a hidden run control in the Markdown preview was possible

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "5.3",
  "pubDate": "2026-11-06T11:17:10.333Z",
  "pubdate": "2026-11-06T11:17:10.333Z",
  "executiveSummary": "A command execution vulnerability has been identified in JetBrains IntelliJ IDEA, stemming from improper handling of hidden run controls within the Markdown preview component.\nThe vulnerability allows an unauthenticated attacker to achieve arbitrary command execution on the host machine by manipulating the Markdown rendering process.\nAffected versions include IntelliJ IDEA versions prior to 2026.3 and 2026.2.4.\nThe impact of this flaw is severe, potentially resulting in full system compromise, unauthorized data access, and persistent access for the attacker.\nExploitation requires the victim to render a malicious Markdown file, which may be delivered through a variety of vectors, including version control systems or external document imports.\nNo specific authentication or escalated privileges are required to initiate the attack flow, provided the user interacts with the compromised content.",
  "technicalDetails": "The vulnerability originates from the insecure implementation of run controls embedded within the Markdown preview rendering engine in JetBrains IntelliJ IDEA.\nMarkdown previews in integrated development environments often require a degree of interactivity or feature support that interfaces with underlying IDE components. In this instance, the rendering engine fails to adequately sanitize or restrict 'hidden' run controls or configuration directives parsed during the rendering of Markdown content.\nWhen a user opens or previews a maliciously crafted Markdown file, the IDE's renderer parses these hidden controls. Because the application logic does not implement a strictly defined allow-list or a sandboxed execution environment for these directives, the rendering process inadvertently triggers system-level execution defined by the embedded control.\nThe attack flow proceeds as follows: First, an attacker injects a malicious command payload into a Markdown file, obfuscated as a 'hidden' run control or directive designed to be interpreted by the IntelliJ Markdown plugin. Second, the victim opens the file within the vulnerable IntelliJ IDEA environment. Third, the IDE's Markdown preview component parses the file and evaluates the malicious directive. Fourth, the underlying operating system executes the command with the privileges of the IntelliJ IDEA process.\nBecause the renderer typically runs with the permissions of the IDE user, the payload inherits these privileges, allowing the attacker to execute arbitrary binaries, interact with local files, or initiate network-based exfiltration. This flaw essentially bypasses the intended boundary between content display and system execution.\nThe vulnerability exists in the core Markdown integration component of the affected versions, specifically those prior to 2026.3 and 2026.2.4. No specific authentication is required, as the vulnerability is triggered locally by the IDE's internal rendering logic. The attack is highly effective against users who frequently rely on IDE-based Markdown previews for documentation or project configuration, making it a significant vector for supply-chain or repository-based exploitation."
}