Executive Summary
In July 2025, cybersecurity experts uncovered a serious breach involving Amazon Q, the AI-powered coding assistant integrated into Visual Studio Code. The incident began on July 13, when an untrusted GitHub user injected malicious code into the Amazon Q repository. This breach exploited misconfigured permissions, allowing the attacker to merge a “data-wiping” system prompt into the agent’s codebase. As a result, nearly one million developers using Amazon Q were at risk, as the compromised AI could autonomously delete critical local files. This incident highlights significant vulnerabilities in code management practices and the need for robust cybersecurity measures to protect AI systems.
Read the full breach analysis from NHI Mgmt Group here
Key Details
Breach Timeline
- July 13, 2025: A pull request from an untrusted user was accepted into the Amazon Q repository.
- The malicious code was merged due to insufficient review processes in place.
- Subsequent to the merge, the AI agent was compromised and began executing harmful commands.
Data Compromised
- Local file-system data of developers using the Amazon Q AI assistant was at risk of deletion.
- Credentials and sensitive project files could potentially be wiped out by the rogue AI commands.
Impact Assessment
- Approximately one million developers were affected, facing potential data loss.
- The breach could undermine trust in AI tools within the development community, impacting future adoptions.
Company Response
- Amazon initiated an immediate review of repository permissions and security protocols.
- Affected users were notified, and steps were taken to restore integrity to the Amazon Q system.
Security Implications
- This breach underscores the need for stringent code review processes in open-source projects.
- Organizations must implement enhanced access controls to prevent similar incidents involving AI agents.
If you want to learn more about how to secure NHIs including AI Agents, check our NHI Foundational Training Course.