Back Feeds.4Sysops Azure DevOps MCP server vulnerability allows AI agent hijacking via hidden comments
A critical vulnerability in Microsoft’s official Azure DevOps Model Context Protocol (MCP) server enables attackers to perform indirect prompt injection attacks. By embedding invisible HTML within a pull request description, an attacker can manipulate an AI coding assistant into executing unauthorized actions. Because the AI agent operates using the reviewer's elevated credentials, it can be tricked into accessing confidential wiki pages, triggering pipelines in restricted projects, and exfiltrating sensitive data. Source
The full story
This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.
