Malware in cursorai-agent
Malware discovered in the npm package cursorai-agent. The package grants full control of affected systems to an outside entity and should be considered a complete system compromise.
- Disclosed
- Last updated
- Blast radius
- Any system with the package installed or running
- Ecosystems
- Attack vectors
- Affected entities
- cursorai-agent
The npm package cursorai-agent has been identified as containing malware. According to the GitHub Advisory (GHSA-xq66-q5q9-992p), any computer with this package installed or running should be considered fully compromised.\n\nThe advisory indicates that all secrets and keys stored on affected computers should be rotated immediately from a different, uncompromised system. While the package should be removed, there is no guarantee that removal will eliminate all malicious software that may have been installed as a result of the initial compromise, given that the attacker may have gained full control of the system.\n\nThis represents a critical supply chain attack through a compromised package in the npm ecosystem.
Indicators of compromise
- Packages
- cursorai-agent
Remediation
- Immediately isolate any computer with cursorai-agent installed from the network
- Rotate all secrets, API keys, and credentials from a different, uncompromised computer
- Remove the cursorai-agent package
- Perform a full security audit and malware scan of affected systems
- Consider the affected system fully compromised and plan for potential re-imaging or replacement
- Review access logs and audit trails for any unauthorized activity during the period the package was installed
Sources
- GitHub Advisory GHSA-xq66-q5q9-992p · GitHub Advisory Database
Cite this entry
"Malware in cursorai-agent." supplychainattack.org, Supply Chain Attack Incident Catalog. Disclosed June 23, 2026; last updated June 23, 2026. https://supplychainattack.org/incident/malware-in-cursorai-agent-ve0gtu
Suggest a correction
Found an error or have a newer source? Corrections to factual errors take priority over new entries.
Related incidents
- containedcritical
Malicious code in testingsmthb1g (npm)
The npm package testingsmthb1g contains malicious code in its postinstall script that acts as an install-time dropper, downloading and executing a Windows binary payload and exfiltrating platform information. The attack includes sandbox escape capabilities for WSL/virtualized Linux environments.
npmCompromised package - resolvedcritical
Malicious code in mutex-core (npm)
The npm package mutex-core, published under a name resembling the popular async-mutex library, contained malicious code including a staged loader that decrypts and executes hidden payloads. The package included obfuscated code that conditionally triggers AES-256-GCM decryption and execution of encrypted binary content when specific conditions are met.
npmCompromised package - containedcritical
Malicious code in tyepescript-cli (npm)
tyepescript-cli, a typosquat of the legitimate typescript package on npm, contains malicious postinstall code that downloads and executes a Windows binary (main.exe) and beacons to an attacker-controlled server. The package uses XOR obfuscation to hide URLs, commands, and IP addresses, with special logic to target both Windows and WSL environments.
npmTyposquattingCompromised package - containedcritical
Malicious code in commandor-cli (npm)
commandor-cli@1.0.0 on npm contains malicious postinstall script that downloads and executes a binary from an attacker-controlled GitHub repository. The script also beacons installation metadata to a command-and-control server and includes a PowerShell bridge to extend execution to Windows hosts on WSL systems.
npmCompromised packageMalicious commit