文章收藏
Building agents: When to fold subagents into your main agent loop
Anthropic 应用 AI 团队：多 Agent 部署中，单 Agent + Skills 常优于编排器加专家子 Agent；给出何时该拆、如何分阶段折叠迁移的实践清单。
How the Claude Code team does code review
Claude Code's team shifted review to machines with layered automated checks and risk-scaled human approval. Merged PRs doubled while reverts fell over 40%.
Redesigning source control at Anthropic
Anthropic's daily merge rate exploded past 30,000 changes, straining git. They built Scion, a client-server version control system for always-online agents, and migrated their monorepo from git in five months.
Using Claude Code: The Unreasonable Effectiveness of HTML
How and why members of the Claude Code team use HTML instead of Markdown to produce richer, more readable, and easily shareable outputs.
MiniMax Agent Team: 为长程任务，持续进化而生
今天我们介绍 MiniMax Agent 的整体升级，我们将升级后的 Agent 起了个新名字：Mavis — MiniMax as a Jarvis，你的 AI 管家。
Harness Design for Long-Running Application Development
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.