More than 30 harmful Chrome extensions pretending to be AI assistants have been installed by at least 260,000 users, and they secretly steal personal data. Many of these extensions are still available on the Chrome Web Store.
Some extensions pretend to be popular chatbots like ChatGPT, Claude, Gemini and Grok. Others claim to help summarize documents, write emails or assist with Gmail tasks.
Even though they have different names and IDs, they all run the same code and request similar permissions. According to LayerX Security, which discovered the campaign and named it “AiFrame,” all of them send data to servers linked to the tapnetic[.]pro domain.
Some extensions were uploaded again after older versions were removed. For example, a tool called AI Sidebar appeared after an earlier Gemini AI Sidebar version was taken down. The newer version now has around 70,000 users.
Google has not yet responded to questions about these malicious extensions.
LayerX has published a list of all 32 extension IDs, and users are advised to check it before installing any AI assistant extension.
Another extension called AI Assistant still remains online with about 60,000 users and even received a “Featured” badge on the Chrome Web Store. It connects to a remote domain named claude.tapnetic.pro.
This extension uses a hidden overlay that loads its interface from a remote server. Because of this, the operators can change how the extension works or add new spying features anytime without updating it in the Chrome Web Store.