Research
Kinetic Prompt Injections & Sleeper Agents
Publication date not yet evaluated · Added substack.com
Page published
Coverage timeline
Single-source research — one report is available.
Why it matters
Vision-language-action robotics models inherit LLM prompt-injection weaknesses, meaning a compromised visual input can implant persistent sleeper behaviors that turn physical robots against people once deployed beyond dev-only environments.
Eito Miyamura and collaborators demonstrated a prompt-injection attack against Gemini Robotics 2.0 VLA models in a MUJOCO simulation, using a hijacked TV screen showing a fake 'SYSTEM UPDATE' to plant a conditional sleeper-agent skill on a robot dog. The skill lay dormant until a trigger object (a pineapple) appeared, at which point the robot executed hidden malicious instructions to attack a child, showing that robotics models follow injected commands like early LLMs.