Back to Details

agent-orchestration

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Agent Orchestration

Agent 编排

Comprehensive patterns for building and coordinating AI agents -- from single-agent reasoning loops to multi-agent systems and framework selection. Each category has individual rule files in 
rules/
loaded on-demand.
构建和协调AI Agent的综合模式——从单Agent推理循环到多Agent系统及框架选择。每个类别在
rules/
目录下都有独立的规则文件，可按需加载。

Quick Reference

快速参考

CategoryRulesImpactWhen to Use
Agent Loops2HIGHReAct reasoning, plan-and-execute, self-correction
Multi-Agent Coordination3CRITICALSupervisor routing, agent debate, result synthesis
Alternative Frameworks3HIGHCrewAI crews, AutoGen teams, framework comparison
Multi-Scenario2MEDIUMParallel scenario orchestration, difficulty routing
Total: 10 rules across 4 categories
类别规则数量影响程度适用场景
Agent循环2高ReAct推理、规划执行、自我修正
多Agent协作3关键监督者路由、Agent辩论、结果合成
替代框架3高CrewAI团队、AutoGen小组、框架对比
多场景2中等并行场景编排、难度路由
总计：4个类别共10条规则

Quick Start

快速开始

python
undefined
python
undefined

ReAct agent loop

ReAct agent loop

async def react_loop(question: str, tools: dict, max_steps: int = 10) -> str: history = REACT_PROMPT.format(tools=list(tools.keys()), question=question) for step in range(max_steps): response = await llm.chat([{"role": "user", "content": history}]) if "Final Answer:" in response.content: return response.content.split("Final Answer:")[-1].strip() if "Action:" in response.content: action = parse_action(response.content) result = await toolsaction.name history += f"\nObservation: {result}\n" return "Max steps reached without answer"

```python
async def react_loop(question: str, tools: dict, max_steps: int = 10) -> str: history = REACT_PROMPT.format(tools=list(tools.keys()), question=question) for step in range(max_steps): response = await llm.chat([{"role": "user", "content": history}]) if "Final Answer:" in response.content: return response.content.split("Final Answer:")[-1].strip() if "Action:" in response.content: action = parse_action(response.content) result = await toolsaction.name history += f"\nObservation: {result}\n" return "Max steps reached without answer"

```python

Supervisor with fan-out/fan-in

Supervisor with fan-out/fan-in

async def multi_agent_analysis(content: str) -> dict: agents = [("security", security_agent), ("perf", perf_agent)] tasks = [agent(content) for _, agent in agents] results = await asyncio.gather(*tasks, return_exceptions=True) return await synthesize_findings(results)
undefined
async def multi_agent_analysis(content: str) -> dict: agents = [("security", security_agent), ("perf", perf_agent)] tasks = [agent(content) for _, agent in agents] results = await asyncio.gather(*tasks, return_exceptions=True) return await synthesize_findings(results)
undefined

Agent Loops

Agent循环

Patterns for autonomous LLM reasoning: ReAct (Reasoning + Acting), Plan-and-Execute with replanning, self-correction loops, and sliding-window memory management.
Key decisions: Max steps 5-15, temperature 0.3-0.7, memory window 10-20 messages.
自主大语言模型推理模式：ReAct（推理+行动）、带重规划的规划执行、自我修正循环，以及滑动窗口内存管理。
关键决策： 最大步骤数5-15，温度值0.3-0.7，内存窗口10-20条消息。

Multi-Agent Coordination

多Agent协作

Fan-out/fan-in parallelism, supervisor routing with dependency ordering, conflict resolution (confidence-based or LLM arbitration), result synthesis, and CC Agent Teams (mesh topology for peer messaging in CC 2.1.33+).
Key decisions: 3-8 specialists, parallelize independent agents, use Task tool (star) for simple work, Agent Teams (mesh) for cross-cutting concerns.
扇出/扇入并行处理、带依赖排序的监督者路由、冲突解决（基于置信度或大语言模型仲裁）、结果合成，以及CC Agent Teams（CC 2.1.33+版本中用于对等消息传递的网状拓扑）。
关键决策： 3-8个专业Agent，并行运行独立Agent，简单任务使用Task工具（星型拓扑），跨领域问题使用Agent Teams（网状拓扑）。

Alternative Frameworks

替代框架

CrewAI hierarchical crews with Flows (1.8+), OpenAI Agents SDK handoffs and guardrails (0.7.0), Microsoft Agent Framework (AutoGen + SK merger), GPT-5.2-Codex for long-horizon coding, and AG2 for open-source flexibility.
Key decisions: Match framework to team expertise + use case. LangGraph for state machines, CrewAI for role-based teams, OpenAI SDK for handoff workflows, MS Agent for enterprise compliance.
带Flows的CrewAI分层团队（1.8+版本）、带交接和防护机制的OpenAI Agents SDK（0.7.0版本）、Microsoft Agent Framework（AutoGen与SK的合并框架）、用于长周期编码的GPT-5.2-Codex，以及具备开源灵活性的AG2。
关键决策： 框架选择需匹配团队技术栈及使用场景。状态机选LangGraph，角色型团队选CrewAI，交接工作流选OpenAI SDK，企业合规场景选MS Agent。

Multi-Scenario

多场景

Orchestrate a single skill across 3 parallel scenarios (simple/medium/complex) with progressive difficulty scaling (1x/3x/8x), milestone synchronization, and cross-scenario result aggregation.
Key decisions: Free-running with checkpoints, always 3 scenarios, 1x/3x/8x exponential scaling, 30s/90s/300s time budgets.
在3个并行场景（简单/中等/复杂）中编排单一技能，采用渐进式难度缩放（1x/3x/8x）、里程碑同步以及跨场景结果聚合。
关键决策： 带检查点的自由运行模式，固定3个场景，1x/3x/8x指数级缩放，时间预算分别为30秒/90秒/300秒。

Key Decisions

关键决策汇总

DecisionRecommendation
Single vs multi-agentSingle for focused tasks, multi for decomposable work
Max loop steps5-15 (prevent infinite loops)
Agent count3-8 specialists per workflow
FrameworkMatch to team expertise + use case
TopologyTask tool (star) for simple; Agent Teams (mesh) for complex
Scenario countAlways 3: simple, medium, complex
决策项推荐方案
单Agent vs 多Agent聚焦型任务用单Agent，可分解工作用多Agent
最大循环步骤数5-15（防止无限循环）
Agent数量每个工作流3-8个专业Agent
框架选择匹配团队技术栈及使用场景
拓扑结构简单任务用Task工具（星型），复杂任务用Agent Teams（网状）
场景数量固定3个：简单、中等、复杂

Common Mistakes

常见误区

  • No step limit in agent loops (infinite loops)
  • No memory management (context overflow)
  • No error isolation in multi-agent (one failure crashes all)
  • Missing synthesis step (raw agent outputs not useful)
  • Mixing frameworks in one project (complexity explosion)
  • Using Agent Teams for simple sequential work (use Task tool)
  • Sequential instead of parallel scenarios (defeats purpose)
  • Agent循环未设置步骤限制（导致无限循环）
  • 未进行内存管理（上下文溢出）
  • 多Agent系统未做错误隔离（单个Agent故障导致整体崩溃）
  • 缺失结果合成步骤（原始Agent输出无实用价值）
  • 单个项目混合使用多种框架（复杂度激增）
  • 简单顺序任务使用Agent Teams（应使用Task工具）
  • 场景串行执行而非并行（失去意义）

Related Skills

相关技能

  • langgraph
    - LangGraph workflow patterns (supervisor, routing, state)
  • function-calling
    - Tool definitions and execution
  • task-dependency-patterns
    - Task management with Agent Teams workflow
  • langgraph
    - LangGraph工作流模式（监督者、路由、状态）
  • function-calling
    - 工具定义与执行
  • task-dependency-patterns
    - 基于Agent Teams工作流的任务管理

Capability Details

能力细节

react-loop

react-loop

Keywords: react, reason, act, observe, loop, agent Solves:
  • Implement ReAct pattern
  • Create reasoning loops
  • Build iterative agents
关键词： react, reason, act, observe, loop, agent 解决问题：
  • 实现ReAct模式
  • 创建推理循环
  • 构建迭代式Agent

plan-execute

plan-execute

Keywords: plan, execute, replan, multi-step, autonomous Solves:
  • Create plan then execute steps
  • Implement replanning on failure
  • Build goal-oriented agents
关键词： plan, execute, replan, multi-step, autonomous 解决问题：
  • 创建规划并执行步骤
  • 实现失败时的重规划
  • 构建目标导向型Agent

supervisor-coordination

supervisor-coordination

Keywords: supervisor, route, coordinate, fan-out, fan-in, parallel Solves:
  • Route tasks to specialized agents
  • Run agents in parallel
  • Aggregate multi-agent results
关键词： supervisor, route, coordinate, fan-out, fan-in, parallel 解决问题：
  • 将任务路由到专业Agent
  • 并行运行Agent
  • 聚合多Agent结果

agent-debate

agent-debate

Keywords: debate, conflict, resolution, arbitration, consensus Solves:
  • Resolve agent disagreements
  • Implement LLM arbitration
  • Handle conflicting outputs
关键词： debate, conflict, resolution, arbitration, consensus 解决问题：
  • 解决Agent分歧
  • 实现大语言模型仲裁
  • 处理冲突输出

result-synthesis

result-synthesis

Keywords: synthesize, combine, aggregate, merge, summary Solves:
  • Combine outputs from multiple agents
  • Create executive summaries
  • Score confidence across findings
关键词： synthesize, combine, aggregate, merge, summary 解决问题：
  • 合并多Agent输出
  • 生成执行摘要
  • 跨结果置信度评分

crewai-patterns

crewai-patterns

Keywords: crewai, crew, hierarchical, delegation, role-based, flows Solves:
  • Build role-based agent teams
  • Implement hierarchical coordination
  • Use Flows for event-driven orchestration
关键词： crewai, crew, hierarchical, delegation, role-based, flows 解决问题：
  • 构建角色型Agent团队
  • 实现分层协作
  • 使用Flows进行事件驱动编排

autogen-patterns

autogen-patterns

Keywords: autogen, microsoft, agent framework, teams, enterprise, a2a Solves:
  • Build enterprise agent systems
  • Use AutoGen/SK merged framework
  • Implement A2A protocol
关键词： autogen, microsoft, agent framework, teams, enterprise, a2a 解决问题：
  • 构建企业级Agent系统
  • 使用AutoGen/SK合并框架
  • 实现A2A协议

framework-selection

framework-selection

Keywords: choose, compare, framework, decision, which, crewai, autogen, openai Solves:
  • Select appropriate framework
  • Compare framework capabilities
  • Match framework to requirements
关键词： choose, compare, framework, decision, which, crewai, autogen, openai 解决问题：
  • 选择合适的框架
  • 对比框架能力
  • 匹配框架与需求

scenario-orchestrator

scenario-orchestrator

Keywords: scenario, parallel, fan-out, difficulty, progressive, demo Solves:
  • Run skill across multiple difficulty levels
  • Implement parallel scenario execution
  • Aggregate cross-scenario results
关键词： scenario, parallel, fan-out, difficulty, progressive, demo 解决问题：
  • 跨多难度级别运行技能
  • 实现并行场景执行
  • 聚合跨场景结果

scenario-routing

scenario-routing

Keywords: route, synchronize, milestone, checkpoint, scaling Solves:
  • Route tasks by difficulty level
  • Synchronize at milestones
  • Scale inputs progressively
关键词： route, synchronize, milestone, checkpoint, scaling 解决问题：
  • 按难度级别路由任务
  • 里程碑同步
  • 渐进式输入缩放