A hacker inserted data-wiping code into Amazon's generative AI assistant, the Q Developer Extension for Visual Studio Code.
Amazon Q is a free tool that helps developers with coding, debugging, documentation, and custom setup tasks using generative AI. It is available on the Visual Studio Code (VSC) marketplace and has close to one million installations.
On July 13, a hacker known as ‘lkmanka58’ managed to push unauthorized code to Amazon Q’s GitHub repository. The added code contained a data wipe prompt instructing the system to reset to near-factory settings and delete files and cloud resources. Although the code was faulty and caused no damage, it delivered a message about the risks of AI in coding.
The breach occurred after the hacker submitted a pull request from an unverified account. Due to what seems to be poor permission management or misconfigured workflows, the request was approved. Amazon unknowingly published the compromised version, 1.84.0, on July 17.
Security researchers flagged the issue on July 23, prompting Amazon to begin an investigation. The next day, AWS released version 1.85.0 with the malicious code removed.
In a statement, AWS confirmed the breach and explained that its security team conducted a detailed forensic review. They found the unauthorized commit targeted command execution in the Q Developer CLI. AWS immediately revoked credentials, cleaned up the codebase, and published the safe version.
According to AWS, the harmful code did not function in user environments due to formatting issues. Still, some reports suggest the code did run, though it caused no damage. Despite the limited impact, the event is being treated as a serious security incident.
Users who installed version 1.84.0 should upgrade to version 1.85.0 immediately, as the affected release has been removed from all platforms.
Found this article interesting? Follow us on X(Twitter) ,Threads and FaceBook to read more exclusive content we post.
