Claude Code, Gemini CLI, GitHub Copilot Agents Vulnerable to Prompt Injection via Comments

Security researchers have identified a prompt injection vulnerability dubbed 'Comment and Control' affecting major AI coding assistants including Claude Code, Gemini CLI, and GitHub Copilot. The attack exploits comments embedded in code files to manipulate AI agent behaviors, potentially allowing attackers to redirect AI actions, exfiltrate data, or inject malicious code suggestions into development workflows. This vulnerability is particularly concerning as it targets the growing trend of AI agents used in software development, CI/CD pipelines, and automated coding tasks. Organizations using these AI tools should implement input validation, restrict AI agent permissions, monitor AI-generated code changes, and consider disabling automatic code execution features until patches are available.

A researcher has disclosed the details of the AI attack method he has named ‘Comment and Control’. The post Claude Code, Gemini CLI, GitHub Copilot Agents Vulnerable to Prompt Injection via Comments appeared first on SecurityWeek .

Security researchers have identified a prompt injection vulnerability dubbed 'Comment and Control' affecting major AI coding assistants including Claude Code, Gemini CLI, and GitHub Copilot. The attack exploits comments embedded in code files to manipulate AI agent behaviors, potentially allowing attackers to redirect AI actions, exfiltrate data, or inject malicious code suggestions into development workflows. This vulnerability is particularly concerning as it targets the growing trend of AI agents used in software development, CI/CD pipelines, and automated coding tasks. Organizations using these AI tools should implement input validation, restrict AI agent permissions, monitor AI-generated code changes, and consider disabling automatic code execution features until patches are available. A researcher has disclosed the details of the AI attack method he has named ‘Comment and Control’. The post Claude Code, Gemini CLI, GitHub Copilot Agents Vulnerable to Prompt Injection via Comments appeared first on SecurityWeek . A researcher has disclosed the details of the AI attack method he has named ‘Comment and Control’. The post Claude Code, Gemini CLI, GitHub Copilot Agents Vulnerable to Prompt Injection via Comments appeared first on SecurityWeek .