Scale and Selection: What Makes Automatic Harness Evolution Work for Visual-Interface Robot Agents

arXiv:2609.39304 · cs.RO, cs.AI · Submitted 2026-09-30 · Read on arXiv

cs.RO, cs.AI

Submitted: 2026-09-30

Updated: 2026-09-30

Project page: https://galaxygeneralrobotics.github.io/astra-policy

Terminology

• Benchmark — 3× in this paper · explained in 2 episodes
• Benchmarking — 3× in this paper · explained in 3 episodes
• Foundation Model — 2× in this paper · explained in 3 episodes
• Foundation Models — 2× in this paper · explained in 7 episodes
• Generalization — 2× in this paper · explained in 13 episodes
• In Conference on Robot Learning — 2× in this paper
• Trace — 2× in this paper · explained in 2 episodes
• Verification — 2× in this paper · explained in 2 episodes
• Abstraction — 1× in this paper · explained in 1 episode
• Action Space — 1× in this paper · explained in 1 episode
• Scaling Laws — 1× in this paper · explained in 4 episodes
• Steering — 1× in this paper · explained in 2 episodes
• Temporal Reasoning — 1× in this paper · explained in 2 episodes

Sources

• Evolve Vision-Language-Action Model into an Agent with On-the-fly Tool-use
• RHO: Your Coding Agent is Secretly a Roboticist
• CaP-X: A Framework for Benchmarking and Improving Coding Agents for Robot Manipulation
• VIA: Visual Interface Agent for Robot Control
• ReKep: Spatio-Temporal Reasoning of Relational Keypoint Constraints for Robotic Manipulation
• What Are We Actually Benchmarking in Robot Manipulation?
• Recursive Harness Self-Improvement
• Meta-Harness: End-to-End Optimization of Model Harnesses
• Guava: Distilling Frontier VLMs into a Compact Agent through a Robotic Manipulation Harness
• ASPIRE: Agentic /Skills Discovery for Robotics
• Rethinking the Evaluation of Harness Evolution for Agents
• ENPIRE: Agentic Robot Policy Self-Improvement in the Real World
• Self-Harness: Harnesses That Improve Themselves
• Harness VLA: Steering Frozen VLAs into Reliable Manipulation Primitives via Memory-Guided Agents
• Beyond Prompts: Measuring and Optimizing LLM Tool-Agent Harnesses

Related papers

• FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
• MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
• RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
• HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
• APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies
• Fine-tuning is Not Enough: A Parallel Framework for Collaborative Imitation and Reinforcement Learning in End-to-end Autonomous Driving