Browsing: MCP Protocol

Researchers have identified a critical MCP protocol vulnerability that allows attackers to spread malicious prompts between internal AI agents, exploiting trust gaps in agentic networks.