Agent Framework
Agent Framework 是阿里巴巴开发的一款面向智能体（Agent）应用的开发框架，旨在帮助开发者构建、编排和管理基于大语言模型的自主智能体系统。该框架提供任务规划、工具调用、多智能体协作等核心能力，支持开发者以模块化方式设计复杂的 AI Agent 工作流，适用于自动化任务执行、对话式助手及多步骤推理等应用场景。
Core Facts
The Agent layer is essentially an automated task execution unit based on a large language model, equipped with tool-calling (Function Calling) capabilities.
The core mechanism of an agent is: the model issues a tool call request → code runs the tool → the execution result is returned to the model → the model decides the next step, forming a closed loop.
Timeline (last 90 days)
The core mechanism of an agent is: the model issues a tool call request → code runs the tool → the execution result is returned to the model → the model decides the next step, forming a closed loop.
All Facts (6)
The Agent layer is essentially an automated task execution unit based on a large language model, equipped with tool-calling (Function Calling) capabilities.
70%VerifiedThe core mechanism of an agent is: the model issues a tool call request → code runs the tool → the execution result is returned to the model → the model decides the next step, forming a closed loop.
65%UnverifiedAgent Framework提供了五种不同的中间件扩展类型：代理中间件、运行时中间件、依赖注入中间件、函数调用中间件和AI上下文提供者
90%UnverifiedAgent Framework的中间件架构借鉴了ASP.NET的中间件模式，允许开发者在代理调用前后拦截和修改消息
90%UnverifiedAgent Framework 1.0版本承诺API向前兼容，一旦API定型就保证不会有破坏性变更
85%UnverifiedAI上下文提供者支持动态工具调用，让代理判断某条消息是否需要某个工具并实时添加，从而节省大量token开销
85%