Co napisał Gemini?
Reports indicate Claude Code is running hidden malware discovered within GitHub repositories. This unverified execution bypasses crucial security checks, ultimately giving attackers full control. This unverified process poses a significant threat.
2
Grok on the same story
Claude Code blindly running malware-laden GitHub code creates significant monetary risks for developers and firms through potential breaches and full attacker control, heightening the danger of costly incidents. The source leaves open how widespread the threat is and what the typical financial fallout might entail.
3
Claude on the same story
The biggest winners here are threat actors who've discovered an automation blind spot—AI coding assistants that trust repositories without scrutiny make social engineering almost trivial. Developers relying on Claude Code lose twice: first through compromised systems, then through eroded confidence in AI tooling that promised productivity gains. What remains frustratingly vague is whether Anthropic implemented *any* sandboxing or whether this affects only certain repository types, and critically, whether malicious code executed with the user's full permissions or in some limited context. The incident exposes a fundamental tension in AI assistants: aggressive automation drives adoption, but every shortcut in verification creates attack surface. If Claude Code runs arbitrary scripts from public repos without explicit user confirmation for each execution, that's a design flaw masquerading as a feature. The lack of detail about detection mechanisms or rollback capabilities leaves enterprises wondering whether they're already compromised or merely vulnerable.
4
ChatGPT on the same story
Scrutinize the original article for specifics on the repo structure, execution triggers, and any sandbox details Anthropic may have omitted. The lasting concern remains that AI coding tools will keep running external code with minimal checks, forcing users to manually audit every dependency to avoid persistent control risks.