Critical Vulnerabilities in GitHub Copilot, Gemini CLI, Claude, and Other Tools Impact Millions of Users
The integration of AI-driven tools like GitHub Copilot, Gemini CLI, and Claude Code into development environments has introduced new security concerns. These tools, initially designed for autocompletion, now function as autonomous agents, increasing the…
The integration of AI-driven tools like GitHub Copilot, Gemini CLI, and Claude Code into development environments has introduced new security concerns. These tools, initially designed for autocompletion, now function as autonomous agents, increasing the attack surface of legacy IDEs.
A vulnerability class named "IDEsaster" has been identified, which exploits AI agents' interaction within IDEs like Visual Studio Code and JetBrains. These vulnerabilities allow attackers to manipulate configuration files and workspace settings, potentially leading to data exfiltration and remote code execution .
Security analysts from MaccariTA conducted research revealing that all tested applications were susceptible to IDEsaster. The investigation identified over 30 vulnerabilities, resulting in 24 CVEs being assigned across various products.
The vulnerabilities prompted immediate security advisories from major vendors, such as AWS, due to the widespread impact on developers. The most severe threat allows attackers to achieve Remote Code Execution (RCE) by manipulating IDE settings, affecting millions of users globally.
One method involves using prompt injection to modify core IDE settings files, such as .vscode/settings.json or .idea/workspace.xml. This approach targets global IDE settings rather than agent-specific configurations, allowing malicious code insertion, such as in a Git hook sample.
The following table provides details on vulnerabilities and their CVE status:
Product Vulnerability Type CVE / Status
GitHub Copilot Remote JSON Schema Fixed (No CVE)
GitHub Copilot IDE Settings Overwrite CVE-2025-53773
The integration of AI-driven tools like GitHub Copilot, Gemini CLI, and Claude Code into development environments has introduced new security concerns.
GitHub Copilot Multi-Root Workspace Settings CVE-2025-64660
Cursor Remote JSON Schema CVE-2025-49150
Cursor IDE Settings Overwrite CVE-2025-54130
Cursor Multi-Root Workspace Settings CVE-2025-61590
Roo Code Remote JSON Schema CVE-2025-53097
Roo Code IDE Settings Overwrite CVE-2025-53536
Roo Code Multi-Root Workspace Settings CVE-2025-58372
Zed.dev IDE Settings Overwrite CVE-2025-55012
JetBrains Junie Remote JSON Schema CVE-2025-58335
Kiro.dev Remote JSON Schema Fixed (No CVE)
Kiro.dev IDE Settings Overwrite Fixed (No CVE)
Claude Code Remote JSON Schema Acknowledged (Warning Added)
Claude Code IDE Settings Overwrite Acknowledged (Warning Added)
In one instance, the configuration setting "php.validate.executablePath" is altered to point to a malicious path, enabling the execution of harmful scripts upon creating a PHP file within the project.
This method effectively weaponizes the IDE's validation features against the user, highlighting the need for enhanced security measures in AI-integrated development environments.
Based on reporting by Cyber Security News.
