Microsoft open-source breach targets AI developers' credentials
Microsoft disabled dozens of GitHub repositories after attackers injected password-stealing malware into tools used with Claude Code, Gemini CLI, VS Code, and Azure development workflows.
Read more
Microsoft temporarily removed or disabled dozens of its open-source GitHub projects while investigating malicious code that could steal passwords and other credentials from developers. Security researchers said many affected repositories were connected to Azure or commonly opened through AI coding environments including Claude Code, Gemini's command-line interface, and VS Code. Microsoft said it notified a small number of customers who may have pulled compromised content and restored some repositories after review. The incident is a high-signal warning for agentic software development: coding assistants can make it easier to discover, install, and execute dependencies, expanding the blast radius of a poisoned repository. It also follows an earlier compromise of Microsoft's Durable Task project, raising questions about supply-chain controls and remediation.
Key details: June 8, 2026, At least 70 repositories disabled, Credential-stealing malware, Claude Code, Gemini CLI, VS Code, Azure tools.
Continue swiping for more AI Brief stories.