1 paper
Yusuke Sano, Takeshi Itoga
Vision-language-action (VLA) agents can execute natural-language instructions, yet deployed systems still lack an operational interface: deciding when the instruction is complete.…