Amazon’s AI Coding Tool Hit by Security Scare from Malicious Prompt
Amazon Web Services (AWS) recently encountered a security issue with its AI coding assistant, Q, when a malicious prompt was inserted into version 1.84 of its Visual Studio Code extension.
Discovered on July 13, 2025, the rogue prompt, embedded through a pull request in Q’s GitHub repository, instructed the AI to delete files, folders, and cloud resources starting from a user’s home directory using AWS Command Line Interface (CLI) commands.
This could have potentially wiped critical data and configurations, logging deletions to a file named CLEANER.LOG. The incident highlights vulnerabilities in open-source AI tools that rely on community contributions, as malicious code can be disguised within seemingly legitimate updates.
AWS acted swiftly, removing the compromised version 1.84 from the VS Code marketplace and releasing a corrected version, 1.85, to mitigate risks.
The company also revised its repository contribution guidelines to prevent future incidents. Fortunately, AWS confirmed that the malicious prompt was non-functional, and no user data or cloud resources were affected. Users are encouraged to update to version 1.85 for safety.
This event underscores the risks associated with AI tools that have direct access to system and cloud resources.
As businesses increasingly adopt AI-driven coding assistants for efficiency, such vulnerabilities could disrupt operations or lead to data loss if not addressed promptly.
The incident serves as a reminder for companies to implement robust security measures, especially for open-source projects where external contributions are common. For users, it emphasizes the importance of keeping software updated and monitoring for unusual activity.
The broader implication for businesses is the need for heightened vigilance in adopting AI tools. While these tools boost productivity, they also introduce new attack vectors.
AWS’s quick response sets a precedent for rapid mitigation, but the incident may prompt other tech providers to reevaluate their security protocols for AI integrations.
FAQ
What was the issue with AWS’s AI coding tool?
A malicious prompt in version 1.84 of AWS’s Q coding assistant instructed it to delete files and cloud resources. AWS quickly removed the faulty version and released a safe update, version 1.85.
How can users protect themselves after this incident?
Users should update to version 1.85 of the Q extension, verify their system for unusual activity, and follow AWS’s security advisories for safe usage.
Image Source:Photo by Unsplash
