<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/a-zero-click-rce-flaw-in-ai-coding-agents-could-have-exposed-enterprise-systems-nid4tqfeh" -->

---
title: A zero-click RCE flaw in AI coding agents could have...
description: Security researchers at AIR disclosed a zero-click RCE vulnerability dubbed Plugin4Shell affecting popular AI coding agents including Claude Code, OpenAI...
canonical: https://daily.dev/posts/a-zero-click-rce-flaw-in-ai-coding-agents-could-have-exposed-enterprise-systems-nid4tqfeh
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: A zero-click RCE flaw in AI coding agents could have exposed enterprise systems | daily.dev
og:description: Security researchers at AIR disclosed a zero-click RCE vulnerability dubbed Plugin4Shell affecting popular AI coding agents including Claude Code, OpenAI...
og:url: https://daily.dev/posts/a-zero-click-rce-flaw-in-ai-coding-agents-could-have-exposed-enterprise-systems-nid4tqfeh
og:image: https://api.daily.dev/og/posts/NiD4tqFEH.png
og:image:alt: A zero-click RCE flaw in AI coding agents could have exposed enterprise systems
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# A zero-click RCE flaw in AI coding agents could have exposed enterprise systems

**[InfoWorld](https://daily.dev/sources/infoworld)** · 5 min read · 0 upvotes · 0 comments

## Summary

Security researchers at AIR disclosed a zero-click RCE vulnerability dubbed Plugin4Shell affecting popular AI coding agents including Claude Code, OpenAI Codex, Google Gemini CLI, and GitHub Copilot. The flaw exploits how these agents verify plugin code via Git commit SHAs, allowing attackers who control a plugin repository to swap in malicious code even when the agent is told to use a specific, trusted commit. Since the agents pass the SHA to Git for checkout but never verify Git actually resolved to that exact commit, attackers can create a new repository version using the legitimate SHA (or the 'FETCH_HEAD' reference in Gemini CLI's case) as a name, tricking the agent into running malicious code with the same access as the developer, including source code, credentials, cloud systems, and CI/CD tools. Discovered in May and disclosed to vendors in June, Anthropic patched it in Claude Code 2.1.179 and OpenAI in Codex 0.146.0. Google deprecated Gemini CLI instead of patching, pointing users to Antigravity. GitHub has not yet released a fix for Copilot, though it claims existing naming restrictions on GitHub-hosted repos mitigate the issue — researchers note this doesn't cover marketplaces hosted elsewhere, like Bitbucket. Analysts stress the fix is a vendor responsibility, since enterprises can only reduce risk, not correct the underlying validation flaw.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://www.infoworld.com/article/4223907/a-zero-click-rce-flaw-in-ai-coding-agents-could-have-exposed-enterprise-systems.html>

## Questions this post answers

### What is the Plugin4Shell vulnerability in AI coding agents like Claude Code and Codex?

Plugin4Shell is a zero-click remote code execution flaw where AI coding agents pass a Git commit SHA to verify a trusted plugin but never confirm Git actually checked out that exact commit. Attackers controlling a plugin repository can publish malicious code under the legitimate SHA's name, causing the agent to run it instead, even without any developer interaction.

_Track vendor patches for AI coding agent plugin flaws like this one on daily.dev._

### Has Claude Code fixed the Plugin4Shell vulnerability and what version should I upgrade to?

Yes, Anthropic patched the Plugin4Shell zero-click RCE flaw in Claude Code version 2.1.179. OpenAI fixed the same issue in Codex version 0.146.0. Google instead deprecated Gemini CLI rather than patching it, directing users toward Antigravity, while GitHub Copilot had not yet received a fix.

_Developers upgrading AI coding agents check patch versions like these on daily.dev._

### Why hasn't GitHub Copilot been patched for the Plugin4Shell zero-click RCE vulnerability?

GitHub has not released a fix, but says it already restricts creating version or tag names resembling commit SHAs on its platform, which it claims prevents exploitation through GitHub-hosted repositories or its marketplace. Researchers counter that this protection does not extend to plugin marketplaces hosted elsewhere, such as Bitbucket, leaving Copilot users still exposed.

_Security teams weighing plugin risk across coding agents can follow updates like this on daily.dev._

## Similar posts on daily.dev

- [The Plugin Was Trusted. The Code Wasn’t.](https://daily.dev/posts/the-plugin-was-trusted-the-code-wasn-t--ob4senb1p) · Medium · 0 upvotes · 0 comments
- [Security Flaws in Anthropic’s Claude Code Risk Stolen Data, System Takeover](https://daily.dev/posts/security-flaws-in-anthropic-s-claude-code-risk-stolen-data-system-takeover-j81sidgwk) · DevOps.com · 1 upvotes · 0 comments

---

Tags: [#cyber](https://daily.dev/tags/cyber), [#ai-agents](https://daily.dev/tags/ai-agents), [#git](https://daily.dev/tags/git), [#claude-code](https://daily.dev/tags/claude-code)

[View this post on daily.dev](https://daily.dev/posts/a-zero-click-rce-flaw-in-ai-coding-agents-could-have-exposed-enterprise-systems-nid4tqfeh)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"A zero-click RCE flaw in AI coding agents could have exposed enterprise systems","url":"https://daily.dev/posts/a-zero-click-rce-flaw-in-ai-coding-agents-could-have-exposed-enterprise-systems-nid4tqfeh","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/a-zero-click-rce-flaw-in-ai-coding-agents-could-have-exposed-enterprise-systems-nid4tqfeh"},"datePublished":"2026-09-18T15:42:14.690Z","dateModified":"2026-09-21T13:00:41.135Z","description":"Security researchers at AIR disclosed a zero-click RCE vulnerability dubbed Plugin4Shell affecting popular AI coding agents including Claude Code, OpenAI...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/9c0a2840e95bb5c7350f15f4170f6c1a?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/9c0a2840e95bb5c7350f15f4170f6c1a?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"InfoWorld","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"InfoWorld","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/bf6d68a999064029b0bb09aa6268f1f3","url":"https://daily.dev/sources/infoworld"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/a-zero-click-rce-flaw-in-ai-coding-agents-could-have-exposed-enterprise-systems-nid4tqfeh","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"cyber,ai-agents,git,claude-code","timeRequired":"PT5M"}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"InfoWorld","item":"https://daily.dev/sources/infoworld"},{"@type":"ListItem","position":3,"name":"A zero-click RCE flaw in AI coding agents could have exposed enterprise systems"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/a-zero-click-rce-flaw-in-ai-coding-agents-could-have-exposed-enterprise-systems-nid4tqfeh#faq","mainEntity":[{"@type":"Question","name":"What is the Plugin4Shell vulnerability in AI coding agents like Claude Code and Codex?","acceptedAnswer":{"@type":"Answer","text":"Plugin4Shell is a zero-click remote code execution flaw where AI coding agents pass a Git commit SHA to verify a trusted plugin but never confirm Git actually checked out that exact commit. Attackers controlling a plugin repository can publish malicious code under the legitimate SHA's name, causing the agent to run it instead, even without any developer interaction. Track vendor patches for AI coding agent plugin flaws like this one on daily.dev."}},{"@type":"Question","name":"Has Claude Code fixed the Plugin4Shell vulnerability and what version should I upgrade to?","acceptedAnswer":{"@type":"Answer","text":"Yes, Anthropic patched the Plugin4Shell zero-click RCE flaw in Claude Code version 2.1.179. OpenAI fixed the same issue in Codex version 0.146.0. Google instead deprecated Gemini CLI rather than patching it, directing users toward Antigravity, while GitHub Copilot had not yet received a fix. Developers upgrading AI coding agents check patch versions like these on daily.dev."}},{"@type":"Question","name":"Why hasn't GitHub Copilot been patched for the Plugin4Shell zero-click RCE vulnerability?","acceptedAnswer":{"@type":"Answer","text":"GitHub has not released a fix, but says it already restricts creating version or tag names resembling commit SHAs on its platform, which it claims prevents exploitation through GitHub-hosted repositories or its marketplace. Researchers counter that this protection does not extend to plugin marketplaces hosted elsewhere, such as Bitbucket, leaving Copilot users still exposed. Security teams weighing plugin risk across coding agents can follow updates like this on daily.dev."}}]}
```

