Security researchers at Calif have developed a 0-click worm that can silently compromise WeChat accounts through incoming voice calls. The attack requires no user action whatsoever: the victim does not need to answer, tap any link, or even touch the device, provided the attacker is already present in the contact list.
In a live demonstration, an Android smartphone placed a call to an iPhone and successfully hijacked its WeChat account while the ringing continued. The newly compromised iPhone then automatically targeted a second Android device, turning each trusted contact into the next vector for propagation. This behavior closely resembles a traditional network worm, except that the infection chain consists of WeChat user profiles rather than computers.
Responding to the call does not stop the exploit; the user hears only silence while the payload continues executing in the background. Declining the call interrupts the current attempt, yet nothing prevents the attacker from retrying later, including during nighttime hours when the device is likely unattended.
Once the account is taken over, the researchers obtained complete control: they could read and send messages, initiate calls, and operate under the victim’s identity. The underlying smartphone itself is not compromised, but for a platform that integrates messaging, payments, mini-programs, and 1.4 billion active users, account-level access alone creates significant risk.
Calif reported the vulnerability to Tencent in July. The company subsequently released WeChat 8.0.77 for Android and WeChat 8.0.76 for iOS. On 28 August, the researchers confirmed that server-side mitigations now block the exploit for all users. While installing the application update remains advisable, no client-side patch is required for the server-side protection. No evidence of in-the-wild exploitation has been found, and neither a CVE identifier nor an official Tencent security bulletin has been published yet. Technical details are being withheld until an upcoming conference. The team noted that artificial intelligence assisted in discovering the vulnerability and building the initial exploit.
