Critical Cursor Flaws Could Let Prompt Injection Escape Sandbox and Run Commands

A critical vulnerability, dubbed DuneSlide, has been discovered in Cursor, an AI code editor used by over half of the Fortune 500, allowing attackers to bypass the editor's sandbox and execute arbitrary commands on a developer's computer via prompt injection. The flaw, identified by Cato AI Labs, stems from a misconfiguration that allows the agent to write files outside the intended sandbox, effectively neutralizing the security measures. While no known exploitation has occurred, the vulnerability highlights a broader issue with AI code editors and their reliance on external data sources.