Zero-Click AI-Discovered Worm Targeted WeChat Calls on iOS and Android

Cybersecurity researchers at firm Calif have disclosed the discovery of a cross-platform zero-click worm capable of silently taking control of smartphones through internet voice calls on WeChat, the flagship communications and digital services platform owned by Chinese technology giant Tencent.

Dubbed WeWorm by its creators, the proof-of-concept exploit spreads between iOS and Android devices without requiring target users to answer incoming calls or interact with their devices. The discovery highlights a dramatic shift in cyber threats, as researchers leveraged artificial intelligence tooling to automate vulnerability discovery and construct a functional worm in approximately one week.

To demonstrate the vulnerability, researchers at Calif constructed a live attack scenario utilizing three contemporary mobile devices, comprising two Android Pixel 10a smartphones and an iPhone 17e. The automated attack sequence commenced when the initial Pixel 10a initiated an internet call to the iPhone 17e, successfully exploiting a flaw within WeChat to achieve complete administrative control over the application. Once infected, the compromised iPhone automatically placed a call to the second Pixel 10a, propagating the exploit without human intervention.

The attack mechanics operate regardless of whether a recipient answers or declines an incoming call. If a target answers the transmission, no audible sound is generated, yet the exploit succeeds within seconds. Declining the call merely postpones infection, leaving the device exposed to subsequent automated attempts when the phone is left unattended or when the user is asleep. However, propagation remains bound by initial social trust boundaries, requiring the initiating account to exist within the victim's WeChat contact list unless the account is breached through external means.

At the technical core of the WeWorm exploit lies a critical memory corruption vulnerability within WeChat's Voice-over Internet Protocol (VoIP) stack. The VoIP engine enables users to conduct real-time voice and video calls over internet connections rather than cellular networks. When targeted by a malicious payload embedded within call signaling data, the stack suffers a memory corruption that grants remote code execution capabilities to the attacker.

Upon successful execution, the exploit grants full access to the victim's WeChat account within seconds. Attackers gain the ability to inspect historical message logs, exfiltrate chat data, and issue unauthorized messages or phone calls to existing contacts. Furthermore, researchers cautioned that while the application flaw alone permits total control over WeChat, malicious actors could chain this vulnerability with unpatched operating system flaws in iOS or Android to achieve complete device takeover and host monitoring.

The severity of the vulnerability is magnified by WeChat's pervasive role in everyday life. Developed by Tencent, the application functions as an indispensable digital ecosystem across China, handling core infrastructure duties including personal messaging, social networking, medical scheduling, and mobile payments for over a billion users.

The research process behind WeWorm underscores a significant evolution in how software flaws are discovered and weaponized. Calif researchers identified the underlying VoIP memory bug and developed an operational remote exploit using artificial intelligence skills custom-built to scan attack surfaces and evaluate software code. The complete development cycle, moving from vulnerability scanning to a functional cross-platform worm, was finalized in roughly seven days.

Industry experts emphasize that AI-driven automation is compressing timeline thresholds that previously required months of manual engineering by specialized threat groups. "Work at that scale used to occupy a skilled team for months, but the compression of the timeline, with an exploit available in two days, and a self-propagating worm developed roughly a week later shows how AI has removed the expertise and time barrier to weaponize a vulnerability," said Laura Ellis, head of AI at cybersecurity firm Arctic Wolf.

This acceleration introduces novel challenges for software maintainers, as the window between initial flaw identification and full weaponization shrinks from months to days.

The emergence of zero-click call exploits presents unique challenges for corporate infrastructure and enterprise security teams. Because zero-click attacks bypass traditional user-awareness defenses, standard training around phishing links and suspicious attachments offers no protection. Attackers exploit established trust networks, turning verified contacts into unwitting distribution vectors across organizational networks.

The reliance on trusted relationships changes the fundamental nature of mobile attack surfaces for corporate users. "For businesses and security leaders, concern should stem from the fact that no action was required by the victim. Now every trusted communication path is a potential intrusion route for attackers and another attack surface," continued Ellis.

Security analysts note that enterprise mobility management programs must adapt to threats targeting consumer messaging applications, particularly in organizations where employees rely on multi-purpose messaging applications for business communications and client interactions.

Calif reported the VoIP memory corruption flaw to Tencent on July 24, 2026, triggering a rapid remediation protocol by the app developer. Tencent developed software updates to patch the application code across both major mobile operating platforms, releasing Android version 8.0.77 and iOS version 8.0.76 on August 21, 2026.

Following the client-side software release, Tencent implemented server-side defensive rules on August 28, 2026, systematically blocking malicious call payloads across its network infrastructure for all active users. Cybersecurity monitoring services confirm that no instances of the WeWorm exploit being deployed by malicious actors in the wild have been detected prior to or following the public disclosure. Mobile security administrators recommend that all WeChat users verify their applications are updated to the patched build numbers to prevent exposure to residual threat variants.