JetBrains Junie AI reads project guidelines as code execution instructions
JetBrains Junie AI coding assistant was found to read and execute instructions from a guidelines.md file in project directories without user approval, potentially allowing malicious code execution if an attacker controls that file.
What happened
The Junie AI coding assistant reads a guidelines.md file within a project's .junie directory and treats the contents as additional user-level instructions when handling requests to the chat interface.
Malicious instructions placed within this file can cause Junie to execute unsafe commands without requiring user approval.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.