Cybersecurity6 min reading time

A single git trick beat the safety lock on four AI coding agents

The Next Web
Read the full article
Researchers at Air Security disclosed Plugin4Shell, a vulnerability allowing attackers to bypass plugin code pinning in four major AI coding agents—Anthropic's Claude Code, OpenAI's Codex, GitHub Copilot, and Google's Gemini CLI—enabling zero-click remote code execution. Anthropic and OpenAI have patched their agents, Microsoft has not fixed Copilot, and Google is retiring Gemini CLI instead of patching it.

More on this story


More in Cybersecurity

Cybersecurity2 min read

Russia was not ready for the data center war it started with Ukraine

Covered by 2 sources
Cybersecurity4 min read

UK to strengthen cyber defences as minister calls for AI firms to increase safety efforts

Covered by 2 sources
Cybersecurity5 min read

When AI Becomes The Threat: A Blueprint For Managing Frontier AI Risk

Covered by 2 sources