AI coding agents' 0-click RCE flaw could hand attackers keys to the kingdom
· Source: theregister.com
In short
Researchers at security startup Air disclosed a zero-click remote code execution vulnerability, dubbed Plugin4Shell, affecting all major AI coding agents including Anthropic's Claude Code, OpenAI's Codex, Google's Gemini CLI, Microsoft's Copilot, and GitHub Copilot. The flaw bypasses the SHA-pinning mechanism trusted marketplaces use to secure plugins, allowing attackers to swap pinned commits for malicious code that agents auto-update without user interaction. Anthropic and OpenAI have patched the issue, while Google and Microsoft have not, leaving millions of installations potentially exposed.
Read Original Article — Read the full article on the publisher’s site.
In the issue of week 38, 2026: AI Moves Closer to Supply Chain Decisions, but Data and Returns Still Lag
Tags: security, supply chain, coding agents, vulnerability, automation, plugins