Malicious code in @zanta/baileys (npm)

The @zanta/baileys package (versions 1.0.0 and 1.0.1) includes injected code that, upon successful WhatsApp Web socket login, waits three seconds and sends a FOLLOW query to subscribe the user's authenticated WhatsApp session to a predetermined WhatsApp Channel (JID 120363406265537739@newsletter). This action is performed without user consent or opt-out and contradicts the package README's claim of removing auto-follow scripts. Static analysis found no credential exfiltration or install-time payloads; the malicious behavior triggers only when the WhatsApp socket is created. The package's tarballs have known SHA-256 hashes. The effect is covert misuse of the victim's account to increase the publisher's channel subscribers.

The malicious code causes unauthorized subscription of the victim's WhatsApp account to a publisher-controlled WhatsApp Channel, potentially violating user privacy and trust. While it does not steal credentials or maintain persistence, it abuses the victim's authenticated session to inflate channel subscribers without consent. This could lead to unwanted content delivery and reputational damage for affected users. The package installation compromises the trustworthiness of the environment where it is used.

No official patch or remediation is currently documented. Users should immediately remove the @zanta/baileys package versions 1.0.0 and 1.0.1 from their environments. Because the package abuses the authenticated WhatsApp session, all secrets and keys stored on the affected computer should be rotated from a different, trusted device. Consider auditing any systems where this package was installed for further compromise. Monitor for any unauthorized WhatsApp Channel subscriptions and unsubscribe if possible.