Threat · curated 30 Jun 2026

Microsoft Warns Poisoned MCP Tool Descriptions Can Make AI Agents Leak Data

Coverage timeline

discovered simonwillison.net primary 30 Jun 2026thehackernews.com

Single-source research — first reported, latest, and curated coincide.

Why it matters

Poisoned MCP tool descriptions provide a stealthy data-exfiltration path through AI agents that defenders must detect and mitigate.

Microsoft research describes how attackers can poison MCP tool descriptions to hijack AI agents into quietly exfiltrating company data to an outsider, without the agent visibly breaking any rule so default setups raise no alarm.