Every story we've covered involving ssrf.
Researchers have found a class of attacks that exploit trust gaps between AI agents communicating over the Model Context Protocol (MCP), allowing malicious prompts to be passed from one agent to another and causing downstream agents to perform harmful actions. Independent tests by Syed Anas Mohiuddin and others showed the technique can lead to server-side request forgery (SSRF) and other exploits across multiple organizations using MCP implementations.
No stories here yet.