THE LIBRARY / 持续积累的阅读档案
资料馆。
传统软件由代码描述,而 Agent 由执行轨迹描述
In software, the code documents the app. In AI, the traces do.
扩展 Managed Agents:将大脑与双手解耦
Scaling Managed Agents: Decoupling the brain from the hands
think 工具:让 Claude 在复杂工具使用场景中停下来思考
The "think" tool: Enabling Claude to stop and think in complex tool use situations
通过 MCP 执行代码:构建更高效的 Agent
Code execution with MCP: Building more efficient agents
借助 Agent Skills,让 Agent 胜任现实世界的任务
Equipping agents for the real world with Agent Skills
超越权限提示:让 Claude Code 更安全、更自主
Beyond permission prompts: making Claude Code more secure and autonomous
我们如何构建 Claude Code 自动模式:更安全地跳过权限确认
How we built Claude Code auto mode: a safer way to skip permissions
用并行运行的 Claude 团队构建 C 编译器
Building a C compiler with a team of parallel Claudes
Claude 开发者平台的高级工具使用能力
Introducing advanced tool use on the Claude Developer Platform
编程 Agent 是黑箱:如何看清它们的内部运行过程
Your coding agents are a black box. Here's how to crack them open.
上下文工程实践:构建 Manus 的经验教训
Context Engineering for AI Agents: Lessons from Building Manus
AutoHarness:通过自动合成代码运行框架改进 LLM Agent(论文摘要)
AutoHarness: improving LLM agents by automatically synthesizing a code harness
没有找到相应资料
试试其他关键词,或清除筛选条件。