Check Point Research: 'Yoba' campaign uses fake GitHub projects posing as ChatGPT and Claude desktop apps, sideloading malware next to signed OpenAI and Anthropic executables to install a crypto miner
Check Point Research published 'Tricks from the Past deployed via signed ChatGPT and Claude binaries' on 27 October 2026.
It says it discovered a campaign it tracks as Yoba in which attackers create fake GitHub projects that pretend to be desktop applications like GPT-5.6 Sol, GPT-6 Astra, and Claude Opus, and that "The malware is placed next to legitimate, signed executables from OpenAI and Anthropic so it can be loaded indirectly, a technique commonly known as DLL sideloading".
Check Point says GitHub.io pages were used to describe the capabilities of the software while the malware was hosted as a GitHub release, that the projects still appear among the top results on Google, most likely due to SEO poisoning, and that the campaign avoids dedicated command-and-control infrastructure by relying on Telegram, Tor, Catbox, and DNS over HTTPS.
It says the actor shows indicators that suggest Russian origins. (source: Check Point Research) The Hacker News reported on 29 October that the downloaded payload is injected into svchost.exe, resulting in the execution of an XMRig cryptocurrency miner. (source: The Hacker News)
- Product
- Fake 'GPT-5.6 Sol', 'GPT-6 Astra' and 'Claude Opus' desktop apps on GitHub (Windows)
- Versions
- n/a — malware campaign
- Exploited in Australia?
- unknown
- Patch to
- Download ChatGPT, Codex and Claude desktop apps only from openai.com and anthropic.com (or your managed software catalogue), never from GitHub releases found through search. A signed executable in a ZIP does not make the bundle safe: block or alert on signed binaries running from user download folders next to unexpected DLLs, and watch for new scheduled tasks and folders under ProgramData.
