CVE-2026-49366: In JetBrains IntelliJ IDEA before 2026

Synopsis

A command injection in JetBrains IntelliJ IDEA lets an attacker run arbitrary commands when a developer triggers filename completion against attacker-crafted input. Exploitation is local and requires the developer to interact with the malicious file or path, but needs no prior authentication; a successful attack runs code with the developer's privileges, giving full read, write, and disruption of anything that user can touch. A patched-image rebuild at 2026.1.1 is available on HarborGuard for affected environments.

HarborGuard Coverage

Exploit Conditions

• Network reachabilityNot required

  The attack vector is local; the attacker needs an existing shell, process, or crafted file on the host where IntelliJ IDEA runs.

• AuthenticationNot required

  No prior authentication to IntelliJ IDEA is required to deliver the malicious filename.

• Victim interactionRequired

  A developer must trigger filename completion against the attacker-controlled path for the injection to fire.

• Attack complexityDetail

  Attack complexity is low: the exploit is reliable once the victim invokes completion, with no race or environmental conditions to satisfy.

Blast Radius

• Runs arbitrary commands with the developer's user privileges on the workstation or build agent.
• Reads source code, SSH keys, cloud credentials, and any other files the developer account can access.
• Modifies source trees, commits, and local artifacts, enabling supply-chain tampering through the developer's identity.
• Can disrupt or terminate the IDE and other processes owned by the user.