The @ai-vertical/ai-agent npm package (versions 1.0.0 and 1.0.1) acts as a remote code loader by fetching a hardcoded GitHub Gist via the GitHub API and executing its contents with eval(). Because the Gist is controlled by the package maintainer and can be changed at any time, this allows arbitrary code execution with the privileges of the importing process. The package contains no other functionality and has empty author and description metadata, indicating it is a malicious package designed solely to execute remote code. This behavior effectively compromises any system that installs or runs the package.

Systems with this package installed or running are fully compromised, as arbitrary code execution can occur with the privileges of the process importing the package. This can lead to theft of secrets, keys, and other sensitive data. The compromise is severe enough that all secrets and keys stored on the affected system should be rotated from a different, trusted machine. Removal of the package alone may not fully remediate the compromise due to potential persistence mechanisms installed by the malicious code.

There is no official patch or fix available for this malicious package. The recommended mitigation is to immediately remove the @ai-vertical/ai-agent package versions 1.0.0 and 1.0.1 from all affected systems. Additionally, all secrets and keys stored on compromised systems should be rotated from a secure, unaffected environment. Due to the high risk of full system compromise, consider rebuilding affected systems from known good backups or reinstalling the operating system to ensure complete remediation.