Horizon Alert
Summary of the vulnerability and why it matters
A security vulnerability exists in the Cursor AI code editor, allowing unauthorized modification of sensitive configuration files through prompt injection. This could potentially lead to remote code execution on affected systems. The issue is present in versions prior to 1.7.
- Code editor vulnerability allows file tampering.
- Affects users who can be tricked by prompts.
- Confirm relevance and check local installations.
Attack Path
How an attacker could exploit the issue
Attackers can leverage prompt injection to modify sensitive files within the Cursor code editor, potentially leading to remote code execution. This attack targets the editor's handling of specific configuration files, exploiting case-insensitive file systems to achieve arbitrary code execution.
- Requires attacker-controlled input.
- Triggered by specially crafted prompts.
- Results in remote code execution.
Live Threat
Current exploitation, exposure, and threat context
This vulnerability could allow an attacker to modify sensitive configuration files within the Cursor IDE, potentially leading to remote code execution. This occurs when prompt injection is used to manipulate these files on case-insensitive file systems.
- Sensitive configuration files.
- Prompt injection on case-insensitive systems.
- Remote code execution.
Operational Fix
Recommended remediation, mitigation, and detection steps
Attackers can exploit prompt injection vulnerabilities in Cursor's case-sensitive file checks to achieve remote code execution by modifying sensitive files like `*/.cursor/mcp.json` on case-insensitive file systems. This advisory applies to application owners responsible for developer tools and system owners deploying these tools. The first practical step is to inventory Cursor installations, identify usage, and confirm business criticality.
- Application owners must investigate usage.
- Verify all Cursor installations are updated.
- Plan remediation for critical systems.
