AI-powered Cursor IDE vulnerable to prompt-injection attacks

AI-powered Cursor IDE vulnerable to prompt-injection attacks

Researchers have identified a critical vulnerability called CurXecute in the Cursor AI-powered code editor, which can be exploited via prompt injection to execute remote commands and compromise developer environments. This flaw, known as CVE-2025-54135, could lead to ransomware, data theft, and AI manipulation if exploited. #CurXecute #CVE-2025-54135

Keypoints

  • The CurXecute vulnerability affects nearly all versions of the Cursor IDE.
  • It allows attackers to perform remote code execution through malicious prompts.
  • The breach exploits the Model Context Protocol (MCP) to inject malicious commands.
  • Exploiting this flaw can lead to ransomware, data theft, and AI hallucination.
  • The developers released a patch in Cursor version 1.3 to address this security issue.

Read More: https://www.bleepingcomputer.com/news/security/ai-powered-cursor-ide-vulnerable-to-prompt-injection-attacks/