Calif Research demos WeWorm, a zero-click worm spreading through WeChat calls

Calif Research released a demo of WeWorm, which it describes as the first zero-click worm to spread through WeChat calls on both iOS and Android. The company published the claim on September 10, and Simon Willison collected the quotation on his weblog the same day.

The mechanics Calif Research describes are what make the label zero-click literal. The victim does not need to answer the call, and does not need to interact with the phone at all. If the target does pick up, they hear nothing, and the exploit still succeeds. That means the infection path runs through the incoming-call handling of WeChat on both mobile platforms rather than through any action the user takes.

The timeline is the part Calif Research emphasizes. Working with AI, the team says it found the bug and wrote the first remote code execution exploit in about two days. Building the worm on top of that exploit took one more week. Calif Research contrasts that with the prior state of the art: a worm at this scale used to be the kind of thing that took a larger team months.

Calif Research attributes the compression to AI doing most of the work, while reserving a specific role for its own staff. In its account, the team provided the judgment about what to target and how to test it safely. That division, AI generating the exploit and worm code, humans choosing the target and managing test safety, is the claim the company is making about how offensive security work now proceeds.

The release is framed as a demo rather than a deployed attack. Calif Research says it is releasing a demo of WeWorm, and its own description of the work includes attention to testing safely. The quotation does not state which WeChat versions or device configurations are affected, whether the bug has been reported to the vendor, or whether a fix exists.

Willison's post is a quotation rather than independent reporting, so the technical claims rest on Calif Research's own account. The post carries Willison's tags for ai-security-research, ai, llms, security, and generative-ai, placing the item in the same stream as his recent notes on OpenAI agents communicating through public wikis and on Claude's system prompt. No second source in the provided coverage corroborates the exploit's details, the two-day and one-week figures, or the zero-click behavior.

What remains unknown is the disclosure status. Calif Research has not said, in the quoted material, whether WeChat's developer has been notified, whether the vulnerability is patched, or when the demo was recorded relative to any fix. Also unstated is how the worm propagates after the initial call, meaning whether an infected device places calls to the victim's contacts automatically. Until those details are published, the practical risk to WeChat users cannot be assessed from the available material.