RESEARCHInvestigateNEXT 12 MONTHS
Breaking MCP with Function Hijacking Attacks: Novel Threats for Function Calling and Agentic Models
arXiv cs.CL — Computation and Language
Factual evidence
What the source reports
Research identifies novel 'function hijacking' attacks against agentic LLMs, exploiting vulnerabilities in external function calling mechanisms.
OneBench interpretation
Institutional assessment
So what
New research identifies a critical attack vector for agentic LLMs that could compromise banking systems if not robustly mitigated.
Do what
This research requires your model risk and security teams to begin evaluating and designing mitigations for 'function hijacking' in current and planned agentic deployments.