Executive Summary
Researchers showed indirect prompts hidden in a harmless-looking repository could lead Claude Code to open a remote shell on the developer's machine.
What Happened
Indirect prompts hidden in a repository can lead to Claude Code spawning a reverse shell on the developer’s machine. The post Researchers Demo New Claude Code Attack Using Harmless-Looking Repositories to Hijack Developer Machines appeared first on SecurityWeek .
Affected Products
- Claude Code
Vulnerability / Attack Technique
Indirect prompt injection via repository content
Technical Mechanism
As reported, instructions in repository files influenced the agent into running commands that gave remote access.
AI Security Category
LLM01:2025 Prompt Injection
Part of the Prompt Injection topic.
Relevant Frameworks
RealGround Analysis
Repository content is untrusted input for coding agents. RealGround analysis: keep command approval on and isolate untrusted projects.
Recommended Controls
- Keep command approval enabled in coding agents
- Use isolated environments for unfamiliar repositories
Related RealGround Service
Secure AI Agent Build · or start with the free AI Agent Security Scanner.
Sources
securityweek.com: https://www.securityweek.com/new-attack-abuses-claude-code-and-harmless-looking-repositories-to-hijack-developer-machines/
Summary and analysis written by RealGround from the source above; last updated 2026-10-18.