Online Skill Learning for Web Agents via State-Grounded Dynamic Retrieval

Jiaxi Li , Ke Deng , Yun Wang , Jingyuan Huang , Yucheng Shi , Qiaoyu Tan , Jin Lu , Ninghao Liu

🏛 Institutions: University of Georgia , Tencent America , New York University , PolyU
📅 Date: June 3, 2026
📑 Publisher: arXiv
💻 Env: Web
🔑 Keywords: SGDR online skill learning dynamic retrieval executable skills WebArena

TLDR

SGDR converts successful WebArena trajectories into verified, executable text-code subprocedures using sliding windows, then retrieves a diverse skill set at every step from both the current page state and task goal rather than once per task. It raises average success to 37.5% with GPT-4.1 and 24.3% with Qwen3-4B, improvements of 3.6 and 2.2 percentage points over the strongest compared baseline.

Open paper Code Report issue