Installs into .claude/skills of the current project.
Are you the author of Skill Space Shooting Autonomous Robot Policy?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/hiyenwong-skill-space-shooting-autonomous-robot-policy)
---
name: skill-space-shooting-autonomous-robot-policy
description: Autonomous robot policy improvement through skill-space shooting that enables robots to improve beyond initial training without human demonstration
version: 1.0.0
tags: [cs.RO, cs.AI, cs.LG, robot-improvement, skill-composition, agentic-systems, foundation-models]
source: arxiv
arxiv_id: 2609.38178v1
utility: 0.85
---
# Skill-Space Shooting for Autonomous Robot Policy Improvement
**Authors:** Zihang Rui, Renhao Wang, Haoxu Huang
**Published:** 2026-09-29
**Categories:** cs.RO, cs.AI, cs.LG
**arXiv:** https://arxiv.org/abs/2609.38178v1
## Summary
Robots deployed in the physical world must be able to improve beyond their initial training as they encounter new situations and failures. For this improvement to scale across tasks, it must make effective use of experience without requiring human demonstration of each correction. Recent agentic systems offer a way to reduce this reliance on human effort by using foundation models to autonomously compose learned behaviors to complete tasks. Yet completing tasks this way does not itself teach a task-improving policy to the robot. This paper introduces skill-space shooting, a method that enables robots to autonomously improve their policies by exploring and composing skills in a learned skill space, allowing continuous improvement without requiring human demonstrations for each new situation.
## Key Contributions
- Introduces skill-space shooting for autonomous robot policy improvement
- Enables robots to improve beyond initial training without human demonstrations
- Leverages foundation models to autonomously compose learned behaviors
- Provides a scalable approach for continuous robot improvement across diverse tasks
## Relevance
This paper addresses a critical challenge in robotics: enabling deployed robots to continuously improve without constant human intervention. The skill-space shooting approach is particularly relevant for researchers working on autonomous robots, lifelong learning, and agentic systems. It provides a practical path toward robots that can adapt to new situations and recover from failures autonomously, which is essential for real-world deployment in dynamic environments.
