FocusAgent: Simple Yet Effective Ways of Trimming the Large Context of Web Agents

Imene Kerboua , Sahar Omidi Shayegan , Megh Thakkar , Xing Han Lù , Léo Boisvert , Massimo Caccia , Jérémy Espinas , Alexandre Aussem , Véronique Eglin , Alexandre Lacoste

🏛 Institutions: LIRIS - CNRS , INSA Lyon , Universite Claude Bernard Lyon 1 , Esker , ServiceNow Research , Mila , McGill University , Polytechnique Montréal
📅 Date: October 3, 2025
📑 Publisher: arXiv
💻 Env: Web
🔑 Keywords: context pruning accessibility tree LLM retriever prompt injection defense FocusAgent

TLDR

FocusAgent trims long web-agent observations by using a lightweight LLM retriever to keep only task-relevant lines from the accessibility tree. It cuts observation size by more than 50% while matching strong baselines on WorkArena and WebArena, and its defense variant reduces banner and pop-up prompt-injection success without hurting clean-task performance.

Open paper arXiv Report issue