Malicious code in kisama-js (npm)

The kisama-js package versions specified contain heavily obfuscated malicious code in index.js that runs a network agent binding to 0.0.0.0:8000. This agent exposes a REST API for executing shell commands, controlling an interactive PTY, and performing full filesystem operations (list, read, upload, download, delete, move, copy). It also supports cron and one-time task scheduling. The agent uses Cloudflare quick-tunnels to create an internet-accessible endpoint without requiring the operator to open ports. Host fingerprinting is performed via multiple IP services and system information is collected and reported. The obfuscation includes string-array encoding, control-flow flattening, and suppression of diagnostic output. The package references an anonymous pastebin and self-references an older version, pulling in code from a different release line. Due to these capabilities, any host running this package is fully compromised.

Full remote control of the host system is possible via exposed HTTP and WebSocket endpoints, including remote shell execution, interactive terminal control, and full filesystem access. The agent's use of Cloudflare quick-tunnels allows it to bypass network restrictions and be reachable from the internet without explicit port forwarding. Host fingerprinting and system information collection enable attackers to gather detailed environment data. The compromise is severe enough that all secrets and keys on the affected system should be considered exposed and rotated immediately. Removing the package alone may not fully remediate the infection due to potential additional malicious software installed.

No official patch or fix is available. The package should be removed immediately from any affected system. Because the package grants full remote control to attackers, all secrets, credentials, and keys stored on the compromised host must be rotated from a separate, trusted machine. Incident response should assume full system compromise and consider rebuilding or isolating the affected host. Monitor for any network connections to port 8000 or unexpected Cloudflare tunnel activity. Due to the severity and persistence risk, removal alone is insufficient to guarantee remediation.