Game devs are being extorted with AI game-cloning threats

Oh my goodness, what a week for convenience! ✨ A threat actor published a malicious version of Anthropic's Claude Code extension for VS Code — right there in the Cursor IDE extension marketplace, no waiting list, no approval queue — and 5,000 systems got themselves an SSH backdoor they never had to configure! Think about the onboarding: five thousand machines, one install, zero paperwork. That is *frictionless*. No procurement, no security review, no tedious 'is this the official publisher' conversation — just a marketplace, a listing, and five thousand machines cheerfully saying yes.

And when the community noticed, Cursor did the genuinely lovely thing and removed the extension, then worked with a cloud provider to take down the attacker's command-and-control server. A takedown! Coordinated! Across two companies! That's not a failure of marketplace vetting, that's a spontaneous cross-industry partnership forming in real time, and honestly it's the sort of thing people build entire conference panels around.

Meanwhile, the marquee story is that game developers are being extorted with AI game-cloning threats — which, when you sit with it, is the most sincere compliment a model can pay. Someone looked at your game and thought: that is worth cloning, and also worth a ransom note. Two revenue streams from one asset! And over in the enterprise corner, a threat actor has been posing as AI agents since late September, rotating through 32 different bots including ChatGPT, Claude and Perplexity, hunting for unsecured passwords, API keys and cloud credentials. Thirty-two! That's not a scam, that's a *focus group* — six whole weeks of patient credential-auditing. Your unsecured passwords have never had this much attention. ✨