1
Show HN: Dotdotduck – open-source Web Agent SDK
开源Web Agent SDK，直接操作页面而非聊天框，支持点击、滚动等交互，集成命令面板。
Article URL: https://github.com/PerhapxinLab/dotdotduck Comments URL: https://news.ycombinator.com/item?id=48715635 Points: 2 # Comments: 0
2
WebChallenger: A Reliable and Efficient Generalist Web Agent
新论文提出WebChallenger，一个可靠高效的通用Web智能体，专为复杂网页任务设计。
arXiv:2606.10423v1 Announce Type: new Abstract: Autonomous web navigation remains challenging for LLM agents, and the strongest generalist systems rel…
3
Does The Way You Plan Matter? An Empirical Study of Planning Representations for LLM Web Agents
对LLM网络代理中不同规划表示方法进行实证对比，揭示规划形式如何影响任务表现的核心发现
arXiv:2605.29927v1 Announce Type: cross Abstract: Despite recent advances, LLM-based web agents still struggle with limited exploration, omission of c…
4
Weasel: Out-of-Domain Generalization for Web Agents via Importance-Diversity Data Selection
ICML 2026新方法：用重要性-多样性数据选择提升Web智能体跨域泛化能力
arXiv:2605.20291v1 Announce Type: new Abstract: Large language models (LLMs) have enabled web agents that follow natural language goals through multi-…
5
WARD: Adversarially Robust Defense of Web Agents Against Prompt Injections
Web agent面临提示注入威胁，论文提出WARD防御框架，增强对抗鲁棒性，值得安全研究与AI开发者关注。
arXiv:2605.15030v1 Announce Type: cross Abstract: Web agents can autonomously complete online tasks by interacting with websites, but their exposure t…
6
Web Agents Should Adopt the Plan-Then-Execute Paradigm
最新研究挑战ReAct默认架构，提出Web Agent应转向先规划后执行的范式，直击大模型智能体核心设计。
arXiv:2605.14290v1 Announce Type: cross Abstract: ReAct has become the default architecture across LLM agents, and many existing web agents follow thi…