Research · curated 10 Aug 2026
When Prompts Control Robots: Prompt Injection Attacks in Multi-Agent Robotic Systems
First reported arxiv.org
Coverage timeline
Single-source research — first reported, latest, and curated coincide.
Why it matters
Prompt injection propagating across LLM-driven multi-agent robots shows how a single compromised input can cascade into unsafe physical actions, a critical concern as autonomous robotic systems adopt LLM planning.
A research paper titled "When Prompts Control Robots" systematically evaluates prompt injection attacks against LLM-based multi-agent robotic systems, testing both direct injections into task instructions and indirect injections through perception modules. The authors demonstrate that injected prompts can induce adversarial physical actions, reduce task completion, and propagate across agents via shared prompt structures.