OneBench
Breaking MCP with Function Hijacking Attacks: Novel Threats for Function Calling and Agentic Models | OneBench: AI Insights