How We Build Open Source with AI Agents

1. How We Build Open Source with AI Agents 徐文彬（闻冰）・ LobeHub 创始人蚂蚁终端体验科技大会

2. Arvin Xu 空 FOUNDER, DESIGN ENGINEER LobeHub 创始 / 设计程师 We are a crew of e/acc design engineers on a mission to forge next-gen human-agent interaction paradigms, while building an open-source, transparent, and symbiotic AI ecosystem_ 我们是

3. AI Workspace for Pro User

4.

5.

6. PART 1PART 2PART 3 Human–AI CollaborationThe Scarcity of CorrectnessAgents in Action AI 与正确性是Agents 如何辅助我们

7. Empowered by Agent Teammate in LobeHub 在 LobeHub 除了我们以外，还有很多 Agent 队友

8. Positioning the Collaboration Between Humans and AI Partners_ AI 在产研中的

9. Positioning the Collaboration Between Humans and AI Partners_ AI 在产研中的

10. PART 2 What is our Role? Where is our Scarcity?

11. VERIFIABILITY IS THE LIMIT from Alperen Keleş, University of Maryland Every piece of software in the world has to be “correct” with respect to some notion of correctness. Before LLMs, our time was split between writing, reading and verifying code. With LLMs, we are able to offload code writing phase to LLMs, but the verification process cannot be offloaded, only pushed up to a different layer. 世界上的每所谓的正确性就是指所创建的软件符合创作者的意图， https://alperenkeles.com/posts/verifiability-is-the-limit

12. Correctness Is Becoming Critical 正确性正在变得愈发重要 Ant-Design Test CI AreWeTurboYet.com

13. TDD: Ensuring Correctness at the Function & Module Level 通过测试驱动开发，为核 • 函数/模块级别的测试驱动开发 • 可直接由AI 成 • 由校验测试例与业务表现

14.

15. BDD + E2E: Capturing Correctness from the Business Perspective 通过

16. PART 3 Agents in Action 辅助我们的 Agents

17. CASE 1: i18n Agent 国际化(i18n)实现：多语持的快速实现案

18. https://github.com/lobehub/lobe-cli-toolbox/tree/master/packages/lobe-i18n

19. Stage1_ CLI 模式

20. Stage1_ CLI 模式

21. Stage1_ CLI 模式 ComfyUI 了我们的在 2024 也引案

22. Stage2_ i18n Workflow

23. Stage2_ i18n Workflow 仓库集成，i18n

24. CASE 2: Local System Agent 通过命令

25. an small issue and collaboration with supply chain

26.

27. gh is underrated gh 还在被持续低估通过 gh 这个命令，我们可以轻易实现对： - issue - pr - action - discussion - codebase - … 乎所有 github 的操作等

28. CASE 3: Claude Code Workflow

29. Leave the refactoring to time 把重构交给时间

30. OUR VISION 未来展望

31. In the past, only big corps had the resources and tech to build complex systems boosting efficiency and creativity. Now, we're witnessing a paradigm shift — Human-AI synergy is giving rise to "Solopreneurs" in the wild. 过去，只有如今，我们正

32. From Big Corporations to Solopreneurs_ 从

33.