AWorld MCP

AWorld：你的世界的代理线束

*“人工智能的下一个前沿是你的专业知识”*

     

中文版 | 自动化 | 进化 | 贡献 |

______________________________________________________________________

General AI often hits a "wall of context"—the nuanced data, workflows, and intuition that define your world. An agent's true power lies not in the model alone, but in its Agent Harness: the framework orchestrating its tools, memory, context, and execution.

这是 A世界论文：一个强大的安全带是不够的。只有当像你这样的专家嵌入宝贵的知识，有效地在墙上建造大门时，真正的人工智能扩展才能解锁。

AWorld是为这一独特目的而设计的平台。我们提供了一套完整的、经过实战考验的安全带，作为专家，将您的知识打造成一支自主代理舰队的秘诀。我们共同超越了人工智能的一般承诺，创建了强大、精确的应用程序 你的 特定领域。

从专业知识到产品

看看当专家知识被编码为可重用时会发生什么 技能。以下创作由AWorld Agent精心策划，展示了我们的核心缩放定律：随着社区贡献更多专业知识，整个生态系统变得更加强大。

自 一个提示视频生成 到 深度搜索工作流，每个例子都将专业知识转化为可重复的生产能力。

这就是今天的可能。想象一下我们将用什么构建 *你的* 专业知识。

Capability Expertise See it in Action Recipe

Create App • Auto-creation by base model • Auto-evaluation by UI Evaluation Skill

View Recipe

Deep Search • Auto-search by Agent Browser Skill

View Recipe

Create Video: Self-Intro • Auto-creation by Remotion Skill • Human evaluation

View Recipe -->

Create Video: Calculus • Auto-creation by Remotion Skill • See full video on Youtube

View Recipe -->

One-Prompt Video: Trig-Identity • Auto-creation by Remotion Skill • See full video on Youtube

View Recipe

One-Prompt Video: Corporate Training • Auto-creation by Remotion Skill • See full video on Youtube

View Recipe

One-Prompt Video: Brand Marketing • Auto-creation by Video Diffusion & Audios Insert Skill • See full video on Youtube

View Recipe

One-Prompt Video: Social Media • Auto-creation by Video Diffusion & Audios Insert Skill • See full video on Youtube

View Recipe

One-Prompt Video: Vtuber • Auto-creation by Video Diffusion & Audio Generator & Video Embedded Skill • See full video on Youtube

View Recipe

您的AWorld CLI之旅

从一个想法到一个进化的自主代理的旅程始于你的指尖。

安装并激活

安装一次，全局配置，随时随地运行。

安装AWorld CLI

git clone https://github.com/inclusionAI/AWorld && cd AWorld

conda create -n aworld_env python=3.11 -y && conda activate aworld_env 

pip install -e . && cd aworld-cli && pip install -e .

配置和启动

cd your working directory

aworld-cli --config

配置后，只需在终端中键入aworld-cli即可开始您的旅程。

或者，您可以通过创建 .env 文件在您的 working directory 使用您的模型和API设置。看 AWorld CLI配置 对于核心变量。

使用AWorld CLI自动创建

AWorld-CLI goes beyond simple scaffolding. It acts as a central brain, the AWorld Agent, which orchestrates a team of specialized sub-agents to build, evaluate, and even evolve other agents autonomously.

这个多智能体系统协同工作，将你的想法变成现实：

Agent NameRole & Core Function

👑 AWorld AgentThe Orchestrator: The central brain that interprets user goals, creates a plan, and delegates tasks to the appropriate sub-agents. It manages the entire workflow from start to finish. 🧑‍💻 DeveloperThe Builder: The master craftsman responsible for writing, debugging, and refactoring code. 🧐 EvaluatorThe Judge: The quality assurance expert. It assesses the Developer's output against objective criteria, providing the critical feedback required for the evolution loop. 🎬 Video DiffusionThe Video Creator: A diffusion-model-based sub-agent (e.g., Kling-V3) that generates videos from text or text+image inputs. 🎤 Audio GeneratorThe Voice Creator: A TTS-model-based sub-agent that converts text input into speech audio. 🖼️ Image GeneratorThe Image Creator: A sub-agent that generates images from text or text+image inputs.

进化循环：构建->评估->进化

想象一下，你问： *“帮助我创建一个UI质量分数高于0.9的英语单词学习迷你应用程序。”*

• 开发者构建：The Developer 使用以下命令分析需求并编写代码（例如HTML） 投射.
• 评审员评委：The Evaluator 使用以下命令检查输出 我们经过验证的技能.
• 循环优化：如果分数低于目标（例如0.9），AWorld会指示开发人员解决评估人员发现的具体问题。此循环将一直持续到满足您的条件为止。

*📹 看到自我进化循环的作用*

无评估，无进化

For an agent to improve, it must first understand what "good" looks like. This evaluation is the core of our autonomous evolution loop, but it's a complex challenge. It ranges from objective tasks with clear metrics (e.g., solving a math problem) to subjective ones requiring human preference. Real-world evolution is further complicated by massive codebases, limited context windows, and the need for precise iteration.

AWorld provides the complete infrastructure to master both evaluation scenarios, turning your expertise into the definitive driving force that steers an agent through the entire evolution loop.

CAST：征服代码复杂性

Agents often fail because of overwhelming code complexity. We built CAST (Code Abstract Syntax Tree) to solve this. Instead of seeing a flat text file, CAST gives the agent an architectural blueprint of the code. This enables:

• 分层导航：立即理解代码结构和目的，而不会迷失在实现细节中。
• 近乎无限的上下文：智能压缩代码，只向代理提供相关信息，打破了上下文窗口的限制。
• 手术代码修改：在完全依赖意识的情况下执行精确的更改，避免“盲”文本替换的笨拙错误。

您作为评估者的专业知识

CAST provides the technical capability for change, but your knowledge provides the direction. AWorld's Shared Skill System makes your expertise the ultimate measure of quality.

Automated Evaluation: Evaluator agent judge performance and identify flaws, setting a clear, objective target for the Developer agent. This creates a powerful synergy: the Evaluator sets the target, and the Developer uses the same knowledge to hit it.

Human Evaluation: For tasks demanding subjective judgment, your intuition is the ceiling. You are the ultimate evaluator. Provide natural language feedback at any stage, and the AWorld agent will interpret it as a high-priority instruction for the next evolutionary cycle.

Whether it's an automated score from a Skill you contributed or your direct manual guidance, in AWorld, precise feedback drives precise evolution.

经过验证的工具：卓越基准

The following top rankings on competitive benchmarks are more than just agent achievements—they are direct validation of the AWorld Harness. They prove our robust, battle-tested infrastructure provides the essential foundation for building state-of-the-art AI.

代理基准测试

Category Achievement Performance Key Innovation Date

🤖 Agent

GAIA Benchmark Excellence

Pass@1: 67.89

Pass@3: 83.49

(109 tasks)

Multi-agent system stability & orchestration

2025/08/06

🧠 Reasoning

IMO 2025 Problem Solving

5/6 problems solved in 6 hours

Multi-agent collaboration beats solo models 2025/07/25

🖼️ Multi-Modal

OSWorld Rank 1st

58.0% Success Rate

The more tools the better? 2025/09/18

🖼️ Multi-Modal

VisualWebArena Rank 1st in September

36.5% Success Rate

Automated tool generation

2025/09/25

🔍 Deep-Search

Xbench Excellence

Pass@1: 51 Pass@3: 61

AWorld has its own context engine: Amni.

2025/10/23

资料综合

*徐增庄、郝炳光、王泽川、温云涛、王茂林等。*

*郝炳光、徐增庄、温云涛、徐信义、杨柳等。*

模型训练

*于承岳、陆思远、庄晨义、王东、吴勤彤等。*

1. FunReason：通过自精化多尺度损失和自动数据精化增强大型语言模型的函数调用。 arxiv，2025年。 纸, 模型

*郝炳光、王茂林、徐增庄、彭存印等。*

*郝炳光、王茂林、徐增庄、陈一成等。*

*谭志文、黄家明、吴勤彤、张宏轩、庄晨义、顾金杰*

*陈继凯、陈龙、王东、甘雷雷、庄晨义、顾金杰*

1. 不要只是微调代理，还要调整环境 arxiv，2025年。 纸

*陆思源、王泽川、张宏轩、吴勤彤、甘雷磊、庄晨毅等。*

元学习

*谢志天、吴勤彤、余承跃、庄晨义、顾金杰*

*何凯文、王志伟、庄晨义、顾金杰*

高级功能

后台子代理执行（v2.0）

非阻塞编排：生成子代理，并在它们在后台执行时继续工作。

from aworld.core.common import ActionModel
from aworld.core.tool.builtin.spawn_subagent_tool import SpawnSubagentTool

# Initialize tool
spawn_tool = SpawnSubagentTool(subagent_manager=your_manager, conf=config)

# 1. Start background task (returns immediately)
action = ActionModel(
    action_name='spawn_background',
    params={
        'name': 'deep_researcher',
        'directive': 'Comprehensive research on quantum computing',
        'task_id': 'research_quantum'
    }
)
obs, reward, _, _, info = await spawn_tool.do_step([action])
task_id = info['task_id']

# 2. Orchestrator continues other work (parallel execution!)
# ... analyze existing data, plan next steps, spawn more tasks ...

# 3. Check status (non-blocking)
check_action = ActionModel(
    action_name='check_task',
    params={'task_id': task_id, 'include_result': False}
)
obs, reward, _, _, info = await spawn_tool.do_step([check_action])
print(f"Status: {info['status']}, Elapsed: {info['elapsed']:.2f}s")

# 4. Wait for completion when ready
wait_action = ActionModel(
    action_name='wait_task',
    params={'task_ids': task_id, 'timeout': 300}
)
await spawn_tool.do_step([wait_action])

# 5. Get result
check_action = ActionModel(
    action_name='check_task',
    params={'task_id': task_id, 'include_result': True}
)
obs, reward, _, _, info = await spawn_tool.do_step([check_action])
result = info['result']

性能效益：编排器工作和子代理执行重叠，实现 max(T_orchestrator, T_subagent) 而不是 T_orchestrator + T_subagent基准测试显示，典型工作负载的执行速度提高了3倍。

文档：参见 docs/代理/并行子代理.md

______________________________________________________________________

贡献

Our roadmap includes expanding our AI for Science & Business initiative, deepening our self-evolution capabilities, and growing our library of community-contributed Skills.

我们热烈欢迎开发人员、研究人员和领域专家加入我们。无论您是在增强框架还是贡献您专业领域的技能，您的工作都是有价值的。

存储库行为更改现在通过OpenSpec进行管理：

• 稳定的能力合同生效 openspec/specs/
• 拟议或飞行中的更改实时生效 openspec/changes/
• 历史上的超级大国笔记现在存在 aworld-cli/internal/docs/superpowers/ 不再是活动的变更管理工作流

如需学术引用或希望联系我们，请使用以下BibTeX条目：

@misc{yu2025aworldorchestratingtrainingrecipe,
      title={AWorld: Orchestrating the Training Recipe for Agentic AI}, 
      author={Chengyue Yu and Siyuan Lu and Chenyi Zhuang and Dong Wang and Qintong Wu and Zongyue Li and Runsheng Gan and Chunfeng Wang and Siqi Hou and Gaochi Huang and Wenlong Yan and Lifeng Hong and Aohui Xue and Yanfeng Wang and Jinjie Gu and David Tsai and Tao Lin},
      year={2025},
      eprint={2508.20404},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2508.20404}, 
}