# Make Code as Policy Great Again: Frontier Agents Write, Call, and Evolve Robot Tools

arXiv: https://arxiv.org/abs/2609.39018 · HTML: https://mlsanity.com/paper/2609.39018/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 43 references that are in the map.
- First posted on arXiv 30 Sep 2026.

## Builds on (most cited first, 26 of 26)

- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [Voyager: An Open-Ended Embodied Agent with Large Language Models](https://mlsanity.com/paper/2305.16291/) (2023)
- [WebArena: A Realistic Web Environment for Building Autonomous Agents](https://mlsanity.com/paper/2307.13854/) (2023)
- [Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware](https://mlsanity.com/paper/2304.13705/) (2023)
- [Code as Policies: Language Model Programs for Embodied Control](https://mlsanity.com/paper/2209.07753/) (2022)
- [Inner Monologue: Embodied Reasoning through Planning with Language Models](https://mlsanity.com/paper/2207.05608/) (2022)
- [VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Models](https://mlsanity.com/paper/2307.05973/) (2023)
- [Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots](https://mlsanity.com/paper/2402.10329/) (2024)
- [ReKep: Spatio-Temporal Reasoning of Relational Keypoint Constraints for Robotic Manipulation](https://mlsanity.com/paper/2409.01652/) (2024)
- [Large Language Models as Tool Makers](https://mlsanity.com/paper/2305.17126/) (2023)
- [Deep Imitation Learning for Complex Manipulation Tasks from Virtual Reality Teleoperation](https://mlsanity.com/paper/1710.04615/) (2017)
- [RT-Trajectory: Robotic Task Generalization via Hindsight Trajectory Sketches](https://mlsanity.com/paper/2311.01977/) (2023)
- [CRAFT: Customizing LLMs by Creating and Retrieving from Specialized Toolsets](https://mlsanity.com/paper/2309.17428/) (2023)
- [MOKA: Open-World Robotic Manipulation through Mark-Based Visual Prompting](https://mlsanity.com/paper/2403.03174/) (2024)
- [RoboDojo: A Unified Sim-and-Real Benchmark for Comprehensive Evaluation of Generalist Robot Manipulation Policies](https://mlsanity.com/paper/2607.04434/) (2026)
- [OSWorld-Human: Benchmarking the Efficiency of Computer-Use Agents](https://mlsanity.com/paper/2506.16042/) (2025)
- [RHO: Your Coding Agent is Secretly a Roboticist](https://mlsanity.com/paper/2606.16458/) (2026)
- [Show-Harness: Just a VLM Agent Can Play Robots](https://mlsanity.com/paper/2609.10522/) (2026)
- [A New Approach to Time-Optimal Path Parameterization based on Reachability Analysis](https://mlsanity.com/paper/1707.07239/) (2017)
- [An Unexpected Robot Policy: Early Evaluations of GPT-6 Astra on RoboDojo and Beyond](https://mlsanity.com/paper/2609.24170/) (2026)
- [Spatial Policy: Guiding Visuomotor Robotic Manipulation with Spatial-Aware Modeling and Reasoning](https://mlsanity.com/paper/2508.15874/) (2025)
- [What's the Move? Hybrid Imitation Learning via Salient Points](https://mlsanity.com/paper/2412.05426/) (2024)
- [RoboStream: Weaving Spatio-Temporal Reasoning with Memory in Vision-Language Models for Robotics](https://mlsanity.com/paper/2603.12939/) (2026)
- [Transferring the Intelligence of VLMs to Robotic Control](https://mlsanity.com/paper/2609.22966/) (2026)
- [InSight: Self-Guided Skill Acquisition via Steerable VLAs](https://mlsanity.com/paper/2606.24884/) (2026)
- [Sketch Interface for Teleoperation of Mobile Manipulator to Enable Intuitive and Intended Operation: A Proof of Concept](https://mlsanity.com/paper/2505.13931/) (2025)

## Built on by (most cited first, 1 of 1)

- [SEAR: Self-Evolving Agentic Robotics](https://mlsanity.com/paper/2610.18792/) (2026)

## Abstract (arXiv)

Frontier models can control robots, but reasoning through every reach, grasp, and retreat makes manipulation slow and token-intensive. We revisit code as policy with a different division of labor: models build executable tools, code handles multi-phase motions, and models decide what to do next. We introduce URAI (Universal Robot-Agent Interface), which couples a programming agent that constructs robot tools with an execution agent that uses them in a feedback loop. The programming agent writes reusable and task-specific tools from task intent and refines them through execution feedback and human guidance. The execution agent selects and parameterizes these tools from current observations; each call runs a complete motion locally before returning control to the agent. Unlike delegating subsequent decisions to a generated program, this design retains model-level decision-making between tool executions. Validated tool revisions persist across episodes without updating foundation-model weights, and a shared GUI and API make the same tools available to humans and agents. Across five RoboDojo tasks and four frozen execution agents, URAI raises aggregate success from 18.0% to 53.0% relative to direct fingertip control, with the largest gain on Swap Blocks; with the same tools, a program written in advance reaches only 24% against 56% for two agents deciding after each call. Three of the four agents also finish episodes 1.3-1.5 times faster with 1.5-1.7 times fewer execution-agent output tokens; DeepSeek-V4-Flash's cost barely changes. We further evaluate URAI on seven real-world AgileX dual-arm tasks, spanning object manipulation, cloth folding, and human-interactive tic-tac-toe. URAI connects the coding and decision-making capabilities of frontier agents, organizing robot control around reusable tools that agents can both invoke and revise.