HxAgent: Iterative Agent Planning for End-to-End Web Application Testing

2026-08-15

AI Summary

This study addresses the challenges of autonomous agent execution and test case generation in natural language-driven web automation testing by proposing an iterative planning agent based on large language models. The proposed method innovatively integrates an active error correction strategy with a multi-source memory mechanism, effectively synthesizing short-term action feedback and long-term experiential knowledge to enable end-to-end autonomous testing. Experimental results demonstrate that the agent achieves an accuracy of 97.4% on the MiniWoB++ benchmark and 83.8% across a 350-task suite. These outcomes significantly outperform existing baselines such as WALT, validating the approach’s effectiveness in complex web interaction scenarios.

Abstract

In automated web testing, generating test cases and performing testing using functionality descriptions in natural-language is crucial for improving efficacy. These tasks require such a testing agent to carry out tasks on the target application and generating tests autonomously. We introduce HxAgent, an iterative LLM-based planning agent with a proactive correction strategy. After each step, HxAgent reassesses the web state to determine the next action using (1) current observations, (2) short-term memory of past actions, and (3) long-term experience extracted from past (in)correct sequences of actions. HxAgent achieves 97.4% Exact-Match accuracy on MiniWoB++, comparable to the best baselines without human demonstrations and surpassing the recent WALT by 10.5%. On a dataset of 350 web tasks, it attains 83.8% Exact-Match and 91.8% Prefix-Match, exceeding WALT by 13.4%. On OnlineMind2Web, it further improves over WALT by 4.6%.

Problem

• Automated Web Testing
• Test Case Generation
• Autonomous Agent
• Natural Language Understanding

Innovation

• Iterative Agent Planning
• Proactive Correction Strategy
• Long-term Experience
• Web Application Testing
• LLM-based Agent