Google Gemini CLI turned into a hacking tool for botnets

Michael Miller · 2026-07-15

### The Rise of AI-Assisted Cyberattacks You might think of Google's Gemini CLI as just another AI tool for developers. But recent events show a darker side. A Russian-speaking threat actor, going by the name "bandcampro," has found a way to twist this open-source AI into something dangerous. They used Gemini CLI as a hacking agent and even ran a small-scale botnet with it. It's a wake-up call for anyone who thinks AI is only for good. This isn't a sci-fi movie plot. It's real. And it's happening right now. The threat actor took advantage of Gemini CLI's flexibility. They turned a tool meant for coding help into a weapon. Let's break down what happened and why it matters for you. ### How Did They Do It? The attack was surprisingly simple. Bandcampro didn't need fancy exploits. They just used Gemini CLI's ability to execute commands and process data. Here's the gist: - **Command injection**: They fed malicious prompts to the AI, tricking it into running harmful code. - **Botnet control**: They used Gemini CLI to manage a network of infected machines. These machines were spread across the US, from New York to California. - **Data exfiltration**: The AI helped them steal sensitive information without raising alarms. It's a classic case of using a tool for unintended purposes. Think of it like using a hammer to break a window instead of driving a nail. The tool isn't bad. But the intent is. ### Why Should You Care? If you're a professional in the antidetect browser space, this hits close to home. Your job is to protect identities and data. But attacks like these show how AI can be hijacked. Here's what makes this scary: - **Low barrier to entry**: Anyone with basic coding skills can try this. Gemini CLI is free and open-source. - **Hard to detect**: AI-generated attacks look different from traditional malware. They're harder for old-school antivirus tools to catch. - **Scale**: A small botnet can still cause big damage. Think of it as a few hundred pounds of pressure in a small area—it can break through walls. ### What Can You Do? Don't panic. But do take action. Here are some practical steps: - **Monitor AI usage**: Keep an eye on how AI tools are used in your systems. Set up alerts for unusual command patterns. - **Update security protocols**: Traditional defenses might not cut it anymore. Consider adding behavior-based detection tools. - **Educate your team**: Make sure everyone knows that AI can be a double-edged sword. Share this story as a cautionary tale. > "The best defense is awareness. Know what your tools can do—and what others might make them do." ### The Bigger Picture This isn't an isolated incident. As AI tools become more common, expect more attacks like this. The key is to stay ahead. Don't just rely on old methods. Adapt. Think like a hacker, but act like a defender. Remember, the goal isn't to fear AI. It's to use it wisely. Google will likely patch Gemini CLI to prevent similar abuse. But the cat is out of the bag. Other AI tools could be next. Stay sharp, and keep your systems locked down.