# From Local Skills to Long-Horizon Web Tasks: Progressive Skill Exploration for LLM Web Agents

**Authors:** Xueyang Feng, Xiaohe Bo, Xu Chen, Quanyu Dai, Feng Wen  
**Lab:** Renmin University · **Published:** 2 Nov 2026  
**Links:** [arXiv](https://arxiv.org/abs/2611.03116) · [PDF](https://arxiv.org/pdf/2611.03116) · [This page](https://omen-bit.github.io/ai-paper-daily/paper/2611.03116.html)

## Key numbers

- **3.9%**: …that SCOPE consistently improves exploration dynamics and achieves 3.9% to 18.6% relative gains over the strongest…
- **18.6%**: …consistently improves exploration dynamics and achieves 3.9% to 18.6% relative gains over the strongest baselines in…

## Abstract

Large language models (LLMs) are driving autonomous agents toward real-world environments, with the Web emerging as one of the most representative interaction scenarios. However, training Web agents with naive exploration remains challenging: successful trajectories are rare, and the sparse terminal rewards make it difficult for agents to obtain useful feedback from failed trajectories. Inspired by skill acquisition theory, we propose Skill-Composed Progressive Exploration (SCOPE), a short-to-long learning framework for long-horizon Web agents that decomposes complex task exploration into a gradual expansion process from local skills to complete task-solving capability. Specifically, SCOPE iteratively explores around local skills: Skill Discovery extracts reusable local skills and failure experiences from historical rollouts; Skill Orchestration composes discovered skills into high-level skill chains; and Skill Expansion leverages failure experiences aligned with the current skill chain to guide the agent toward exploring new subskills. Through this process, SCOPE progressively constructs task-completing skill chains across rollouts and internalizes them into the policy parameters through agentic on-policy distillation. Extensive experiments on WebArena, WebVoyager, and WebShop show that SCOPE consistently improves exploration dynamics and achieves 3.9% to 18.6% relative gains over the strongest baselines in each setting.

---

Notes written by AI from the paper, numbers copied from the abstract. Check the paper before you rely on them. From AI Papers Daily.