Building effective agents
构建有效的代理

Over the past year, we've worked with dozens of teams building large language model (LLM) agents across industries. Consistently, the most successful implementations weren't using complex frameworks or specialized libraries. Instead, they were building with simple, composable patterns.
在过去的一年里，我们与多个行业的数十个团队合作，开发了跨领域的大型语言模型（LLM）代理。一致地，最成功的实施方案并未依赖复杂的框架或专用库，而是采用了简单、可组合的模式进行构建。

In this post, we share what we’ve learned from working with our customers and building agents ourselves, and give practical advice for developers on building effective agents.
在这篇文章中，我们分享了与客户合作及自行构建代理过程中所学到的经验，并为开发者提供了构建高效代理的实用建议。

What are agents?  什么是代理？

"Agent" can be defined in several ways. Some customers define agents as fully autonomous systems that operate independently over extended periods, using various tools to accomplish complex tasks. Others use the term to describe more prescriptive implementations that follow predefined workflows. At Anthropic, we categorize all these variations as agentic systems, but draw an important architectural distinction between workflows and agents:
“Agent”可以有多种定义方式。一些客户将 agent 定义为完全自主的系统，它们能够在较长时间内独立运作，利用各种工具完成复杂任务。另一些人则用该术语来描述遵循预定义工作流的更为规范的实现。在 Anthropic，我们将所有这些变体归类为 agentic 系统，但在架构上对工作流和 agent 进行了重要区分：

• Workflows are systems where LLMs and tools are orchestrated through predefined code paths.
  工作流是系统，其中LLMs和工具通过预定义的代码路径进行编排。
• Agents, on the other hand, are systems where LLMs dynamically direct their own processes and tool usage, maintaining control over how they accomplish tasks.
  另一方面，代理是系统，其中LLMs能够动态地指导自身进程和工具的使用，保持对任务完成方式的控制。

Below, we will explore both types of agentic systems in detail. In Appendix 1 (“Agents in Practice”), we describe two domains where customers have found particular value in using these kinds of systems.
以下，我们将详细探讨这两类代理系统。在附录 1（“实践中的代理”）中，我们描述了两个领域，客户在这些领域中发现了使用此类系统的特殊价值。

When (and when not) to use agents
何时（以及何时不）使用代理

When building applications with LLMs, we recommend finding the simplest solution possible, and only increasing complexity when needed. This might mean not building agentic systems at all. Agentic systems often trade latency and cost for better task performance, and you should consider when this tradeoff makes sense.
在构建使用LLMs的应用程序时，我们建议寻找尽可能简单的解决方案，仅在必要时才增加复杂性。这可能意味着根本不构建代理系统。代理系统通常以延迟和成本为代价换取更好的任务执行效果，您应考虑何时这种权衡是合理的。

When more complexity is warranted, workflows offer predictability and consistency for well-defined tasks, whereas agents are the better option when flexibility and model-driven decision-making are needed at scale. For many applications, however, optimizing single LLM calls with retrieval and in-context examples is usually enough.
当需要更多复杂性时，工作流为定义明确的任务提供了可预测性和一致性，而代理则是在大规模需要灵活性和模型驱动决策时更佳的选择。然而，对于许多应用来说，通过检索和上下文示例优化单个LLM调用通常已经足够。

When and how to use frameworks
何时以及如何使用框架

There are many frameworks that make agentic systems easier to implement, including:
有许多框架使得实现代理系统变得更加容易，其中包括：

• LangGraph from LangChain;
  LangGraph 来自 LangChain；
• Amazon Bedrock's AI Agent framework;
  Amazon Bedrock 的 AI 代理框架；
• Rivet, a drag and drop GUI LLM workflow builder; and
  Rivet，一款拖放式 GUI LLM 工作流构建器；
• Vellum, another GUI tool for building and testing complex workflows.
  Vellum，另一款用于构建和测试复杂工作流的图形用户界面工具。

These frameworks make it easy to get started by simplifying standard low-level tasks like calling LLMs, defining and parsing tools, and chaining calls together. However, they often create extra layers of abstraction that can obscure the underlying prompts ​​and responses, making them harder to debug. They can also make it tempting to add complexity when a simpler setup would suffice.
这些框架通过简化诸如调用LLMs、定义和解析工具以及串联调用等标准低级任务，使得入门变得容易。然而，它们常常引入额外的抽象层，这可能会掩盖底层提示和响应，从而增加调试难度。此外，它们还可能诱使人们在简单设置即可满足需求时，增加不必要的复杂性。

We suggest that developers start by using LLM APIs directly: many patterns can be implemented in a few lines of code. If you do use a framework, ensure you understand the underlying code. Incorrect assumptions about what's under the hood are a common source of customer error.
我们建议开发者首先直接使用LLM API：许多模式只需几行代码即可实现。若使用框架，请确保理解其底层代码。对内部机制的错误假设是客户错误的常见来源。

See our cookbook for some sample implementations.
查看我们的食谱，了解一些示例实现。

Building blocks, workflows, and agents
构建模块、工作流程和代理

In this section, we’ll explore the common patterns for agentic systems we’ve seen in production. We'll start with our foundational building block—the augmented LLM—and progressively increase complexity, from simple compositional workflows to autonomous agents.
在本节中，我们将探讨在生产环境中常见的代理系统模式。我们将从基础构建模块——增强型LLM开始，逐步增加复杂性，从简单的组合工作流到自主代理。

Building block: The augmented LLM
构建模块：增强的LLM

The basic building block of agentic systems is an LLM enhanced with augmentations such as retrieval, tools, and memory. Our current models can actively use these capabilities—generating their own search queries, selecting appropriate tools, and determining what information to retain.
代理系统的基础构建模块是一个LLM，它通过增强功能如检索、工具和记忆来提升。我们当前的模型能够主动运用这些能力——生成自身的搜索查询，选择合适的工具，并决定保留哪些信息。

We recommend focusing on two key aspects of the implementation: tailoring these capabilities to your specific use case and ensuring they provide an easy, well-documented interface for your LLM. While there are many ways to implement these augmentations, one approach is through our recently released Model Context Protocol, which allows developers to integrate with a growing ecosystem of third-party tools with a simple client implementation.
我们建议在实施过程中重点关注两个关键方面：将这些功能定制到您的特定用例中，并确保它们为您的LLM提供一个易于使用且文档齐全的接口。虽然实现这些增强功能的方法有很多，但一种方法是采用我们最近发布的模型上下文协议，该协议允许开发者通过简单的客户端实现与不断扩展的第三方工具生态系统进行集成。

For the remainder of this post, we'll assume each LLM call has access to these augmented capabilities.
在本文的其余部分，我们将假设每个LLM调用都能访问这些增强功能。

Workflow: Prompt chaining
工作流程：提示链

Prompt chaining decomposes a task into a sequence of steps, where each LLM call processes the output of the previous one. You can add programmatic checks (see "gate” in the diagram below) on any intermediate steps to ensure that the process is still on track.
提示链将任务分解为一系列步骤，每个LLM调用处理前一个步骤的输出。您可以在任何中间步骤添加程序化检查（参见下图中的“门”），以确保流程仍在正轨上。

When to use this workflow: This workflow is ideal for situations where the task can be easily and cleanly decomposed into fixed subtasks. The main goal is to trade off latency for higher accuracy, by making each LLM call an easier task.
何时使用此工作流程：此工作流程适用于任务能够被轻松且清晰地分解为固定子任务的情况。其主要目的是通过使每个LLM调用成为更简单的任务，来权衡延迟以换取更高的准确性。

Examples where prompt chaining is useful:
提示链在以下场景中非常有用：

• Generating Marketing copy, then translating it into a different language.
  生成营销文案，然后将其翻译成另一种语言。
• Writing an outline of a document, checking that the outline meets certain criteria, then writing the document based on the outline.
  编写文档大纲，检查大纲是否符合特定标准，然后根据大纲撰写文档。

Workflow: Routing  工作流程：路由

Routing classifies an input and directs it to a specialized followup task. This workflow allows for separation of concerns, and building more specialized prompts. Without this workflow, optimizing for one kind of input can hurt performance on other inputs.
路由将输入分类并引导至专门的后续任务。这一工作流程实现了关注点的分离，并构建了更专业的提示。若无此工作流程，针对一种输入类型的优化可能会损害其他输入类型的性能。

When to use this workflow: Routing works well for complex tasks where there are distinct categories that are better handled separately, and where classification can be handled accurately, either by an LLM or a more traditional classification model/algorithm.
何时使用此工作流程：当任务复杂且存在明确分类，最好分别处理时，路由效果良好。分类可以由LLM或更传统的分类模型/算法准确完成。

Examples where routing is useful:
路由有用的示例：

• Directing different types of customer service queries (general questions, refund requests, technical support) into different downstream processes, prompts, and tools.
  将不同类型的客户服务查询（如一般问题、退款请求、技术支持）引导至不同的下游流程、提示和工具。
• Routing easy/common questions to smaller models like Claude 3.5 Haiku and hard/unusual questions to more capable models like Claude 3.5 Sonnet to optimize cost and speed.
  将简单/常见问题路由至像 Claude 3.5 Haiku 这样的较小模型，而将复杂/不常见问题路由至像 Claude 3.5 Sonnet 这样更强大的模型，以优化成本和速度。

Workflow: Parallelization
工作流程：并行化

LLMs can sometimes work simultaneously on a task and have their outputs aggregated programmatically. This workflow, parallelization, manifests in two key variations:
LLMs 有时可以同时处理一项任务，并通过编程方式聚合其输出结果。这种并行工作流程主要表现为两种关键形式：

• Sectioning: Breaking a task into independent subtasks run in parallel.
  分段：将任务分解为可独立运行的子任务并行执行。
• Voting: Running the same task multiple times to get diverse outputs.
  投票：多次运行同一任务以获取多样化的输出。

When to use this workflow: Parallelization is effective when the divided subtasks can be parallelized for speed, or when multiple perspectives or attempts are needed for higher confidence results. For complex tasks with multiple considerations, LLMs generally perform better when each consideration is handled by a separate LLM call, allowing focused attention on each specific aspect.
何时使用此工作流程：当分割后的子任务能够并行以提高速度，或需要多个视角或尝试以获得更高置信度的结果时，并行化是有效的。对于涉及多重考虑的复杂任务，通常在每个考虑因素由单独的LLM调用处理时，LLMs表现更佳，从而使每个具体方面都能得到专注的关注。

Examples where parallelization is useful:
并行化有用的示例：

• Sectioning:  分节：
  • Implementing guardrails where one model instance processes user queries while another screens them for inappropriate content or requests. This tends to perform better than having the same LLM call handle both guardrails and the core response.
    在一个模型实例处理用户查询的同时，另一个模型实例对其进行筛选，以检测不当内容或请求。这种方式通常比让同一个LLM调用同时处理防护栏和核心响应的表现更佳。
  • Automating evals for evaluating LLM performance, where each LLM call evaluates a different aspect of the model’s performance on a given prompt.
    自动化评估LLM性能的评估过程，其中每个LLM调用针对给定提示评估模型性能的不同方面。
• Voting:  投票：
  • Reviewing a piece of code for vulnerabilities, where several different prompts review and flag the code if they find a problem.
    审查一段代码以查找漏洞，其中多个不同的提示会审查代码并在发现问题时标记出来。
  • Evaluating whether a given piece of content is inappropriate, with multiple prompts evaluating different aspects or requiring different vote thresholds to balance false positives and negatives.
    评估给定内容是否不当，通过多个提示评估不同方面或要求不同的投票阈值来平衡误报和漏报。

Workflow: Orchestrator-workers
工作流程：协调器-工作者

In the orchestrator-workers workflow, a central LLM dynamically breaks down tasks, delegates them to worker LLMs, and synthesizes their results.
在协调者-工作者工作流中，中央LLM动态分解任务，将其委派给工作者LLMs，并综合它们的结果。

When to use this workflow: This workflow is well-suited for complex tasks where you can’t predict the subtasks needed (in coding, for example, the number of files that need to be changed and the nature of the change in each file likely depend on the task). Whereas it’s topographically similar, the key difference from parallelization is its flexibility—subtasks aren't pre-defined, but determined by the orchestrator based on the specific input.
何时使用此工作流程：此工作流程非常适合于复杂任务，在这些任务中你无法预知所需的子任务（例如，在编程中，需要更改的文件数量以及每个文件中更改的性质很可能取决于任务本身）。尽管它在拓扑结构上与并行化相似，但关键区别在于其灵活性——子任务并非预先定义，而是由协调器根据具体输入动态决定。

Example where orchestrator-workers is useful:
orchestrator-workers 有用的示例：

• Coding products that make complex changes to multiple files each time.
  每次对多个文件进行复杂更改的编码产品。
• Search tasks that involve gathering and analyzing information from multiple sources for possible relevant information.
  搜索涉及从多个来源收集和分析信息以寻找可能相关信息的任务。

Workflow: Evaluator-optimizer
工作流程：评估者-优化器

In the evaluator-optimizer workflow, one LLM call generates a response while another provides evaluation and feedback in a loop.
在评估者-优化者工作流程中，一个LLM调用生成响应，而另一个则在循环中提供评估和反馈。

When to use this workflow: This workflow is particularly effective when we have clear evaluation criteria, and when iterative refinement provides measurable value. The two signs of good fit are, first, that LLM responses can be demonstrably improved when a human articulates their feedback; and second, that the LLM can provide such feedback. This is analogous to the iterative writing process a human writer might go through when producing a polished document.
何时使用此工作流程：当具备明确的评估标准，且迭代优化能带来可衡量的价值时，此工作流程尤为有效。适合使用的两个标志是：首先，当人类明确表达其反馈时，LLM 的回应能得到显著提升；其次，LLM 能够提供此类反馈。这类似于人类作者在撰写精炼文档时所经历的迭代写作过程。

Examples where evaluator-optimizer is useful:
评估者-优化器有用的场景示例：

• Literary translation where there are nuances that the translator LLM might not capture initially, but where an evaluator LLM can provide useful critiques.
  文学翻译中存在一些细微差别，译者LLM可能最初未能捕捉到，但评估者LLM能够提供有价值的批评意见。
• Complex search tasks that require multiple rounds of searching and analysis to gather comprehensive information, where the evaluator decides whether further searches are warranted.
  需要多轮搜索和分析以收集全面信息的复杂搜索任务，由评估者决定是否需要进一步搜索。

Agents  特工

Agents are emerging in production as LLMs mature in key capabilities—understanding complex inputs, engaging in reasoning and planning, using tools reliably, and recovering from errors. Agents begin their work with either a command from, or interactive discussion with, the human user. Once the task is clear, agents plan and operate independently, potentially returning to the human for further information or judgement. During execution, it's crucial for the agents to gain “ground truth” from the environment at each step (such as tool call results or code execution) to assess its progress. Agents can then pause for human feedback at checkpoints or when encountering blockers. The task often terminates upon completion, but it’s also common to include stopping conditions (such as a maximum number of iterations) to maintain control.
代理正在生产中崭露头角，随着LLMs在关键能力上的成熟——理解复杂输入、进行推理和规划、可靠地使用工具以及从错误中恢复。代理的工作始于人类用户的指令或互动讨论。一旦任务明确，代理便能独立规划并执行，必要时返回人类获取更多信息或判断。在执行过程中，代理每一步都需要从环境中获取“真实情况”（如工具调用结果或代码执行情况），以评估进展。代理可在检查点或遇到障碍时暂停，寻求人类反馈。任务通常在完成时终止，但也会设置停止条件（如最大迭代次数）以保持控制。

Agents can handle sophisticated tasks, but their implementation is often straightforward. They are typically just LLMs using tools based on environmental feedback in a loop. It is therefore crucial to design toolsets and their documentation clearly and thoughtfully. We expand on best practices for tool development in Appendix 2 ("Prompt Engineering your Tools").
代理能够处理复杂任务，但其实现通常较为直接。它们通常只是LLMs在循环中利用基于环境反馈的工具。因此，清晰且周到地设计工具集及其文档至关重要。我们在附录 2（“为您的工具进行提示工程”）中详细阐述了工具开发的最佳实践。

When to use agents: Agents can be used for open-ended problems where it’s difficult or impossible to predict the required number of steps, and where you can’t hardcode a fixed path. The LLM will potentially operate for many turns, and you must have some level of trust in its decision-making. Agents' autonomy makes them ideal for scaling tasks in trusted environments.
何时使用代理：代理适用于那些开放式问题，即难以或无法预测所需步骤数量，且无法硬编码固定路径的场景。LLM 可能会持续运行多个回合，因此你必须对其决策能力有一定程度的信任。代理的自主性使其在可信环境中扩展任务时成为理想选择。

The autonomous nature of agents means higher costs, and the potential for compounding errors. We recommend extensive testing in sandboxed environments, along with the appropriate guardrails.
代理的自主性意味着更高的成本，以及错误累积的潜在风险。我们建议在沙盒环境中进行广泛测试，并配备适当的防护措施。

Examples where agents are useful:
代理有用的例子：

The following examples are from our own implementations:
以下示例来自我们自己的实现：

• A coding Agent to resolve SWE-bench tasks, which involve edits to many files based on a task description;
  一个用于解决 SWE-bench 任务的编码代理，这些任务涉及根据任务描述对多个文件进行编辑；
• Our “computer use” reference implementation, where Claude uses a computer to accomplish tasks.
  我们的“计算机使用”参考实现，其中 Claude 利用计算机完成任务。

Combining and customizing these patterns
结合并定制这些模式

These building blocks aren't prescriptive. They're common patterns that developers can shape and combine to fit different use cases. The key to success, as with any LLM features, is measuring performance and iterating on implementations. To repeat: you should consider adding complexity only when it demonstrably improves outcomes.
这些构建模块并非规定性的，而是开发者可以塑造和组合以适应不同使用场景的常见模式。与任何LLM功能一样，成功的关键在于衡量性能并对实现进行迭代优化。再次强调：只有在明显改善结果时，才应考虑增加复杂性。

Summary  摘要

Success in the LLM space isn't about building the most sophisticated system. It's about building the right system for your needs. Start with simple prompts, optimize them with comprehensive evaluation, and add multi-step agentic systems only when simpler solutions fall short.
在LLM领域取得成功，不在于构建最复杂的系统，而在于打造适合你需求的系统。从简单的提示开始，通过全面的评估进行优化，仅在简单解决方案不足时才引入多步骤的代理系统。

When implementing agents, we try to follow three core principles:
在实现代理时，我们力求遵循三个核心原则：

1. Maintain simplicity in your agent's design.
   在设计代理时保持简洁。
2. Prioritize transparency by explicitly showing the agent’s planning steps.
   通过明确展示代理的规划步骤，优先考虑透明度。
3. Carefully craft your agent-computer interface (ACI) through thorough tool documentation and testing.
   通过详尽的工具文档和测试，精心打造您的代理-计算机接口（ACI）。

Frameworks can help you get started quickly, but don't hesitate to reduce abstraction layers and build with basic components as you move to production. By following these principles, you can create agents that are not only powerful but also reliable, maintainable, and trusted by their users.
框架能助你快速起步，但进入生产阶段时，不妨减少抽象层次，转而使用基础组件构建。遵循这些原则，你不仅能打造出功能强大的代理，还能确保其可靠、可维护，并赢得用户的信赖。

Acknowledgements  致谢

Written by Erik Schluntz and Barry Zhang. This work draws upon our experiences building agents at Anthropic and the valuable insights shared by our customers, for which we're deeply grateful.
由 Erik Schluntz 和 Barry Zhang 撰写。本文基于我们在 Anthropic 构建代理的经验，并得益于客户分享的宝贵见解，对此我们深表感谢。

Appendix 1: Agents in practice
附录 1：实践中的代理

Our work with customers has revealed two particularly promising applications for AI agents that demonstrate the practical value of the patterns discussed above. Both applications illustrate how agents add the most value for tasks that require both conversation and action, have clear success criteria, enable feedback loops, and integrate meaningful human oversight.
我们与客户的合作揭示了 AI 代理的两个特别有前景的应用，这些应用展示了上述模式在实际中的价值。这两个应用都说明了代理如何在需要对话与行动、具备明确成功标准、支持反馈循环并整合有意义的人类监督的任务中，发挥最大的价值。

A. Customer support  A. 客户支持

Customer support combines familiar chatbot interfaces with enhanced capabilities through tool integration. This is a natural fit for more open-ended agents because:
客户支持将熟悉的聊天机器人界面与通过工具集成增强的功能相结合。这对于更开放式的代理来说是一个自然契合，因为：

• Support interactions naturally follow a conversation flow while requiring access to external information and actions;
  支持的交互自然地遵循对话流程，同时需要访问外部信息和执行操作；
• Tools can be integrated to pull customer data, order history, and knowledge base articles;
  工具可以集成以提取客户数据、订单历史记录和知识库文章；
• Actions such as issuing refunds or updating tickets can be handled programmatically; and
  诸如发放退款或更新票务等操作可以通过编程方式处理；
• Success can be clearly measured through user-defined resolutions.
  成功可以通过用户定义的目标清晰地衡量。

Several companies have demonstrated the viability of this approach through usage-based pricing models that charge only for successful resolutions, showing confidence in their agents' effectiveness.
多家公司通过采用基于使用量的定价模式，仅对成功解决的问题收费，展示了这一方法的可行性，彰显了对其代理效能的信心。

B. Coding agents  B. 编码剂

The software development space has shown remarkable potential for LLM features, with capabilities evolving from code completion to autonomous problem-solving. Agents are particularly effective because:
软件开发领域展现了显著的潜力，具备从代码补全到自主问题解决的多样化功能。代理尤其高效，原因如下：

• Code solutions are verifiable through automated tests;
  代码解决方案可通过自动化测试进行验证；
• Agents can iterate on solutions using test results as feedback;
  代理可以通过测试结果作为反馈来迭代解决方案；
• The problem space is well-defined and structured; and
  问题空间定义明确且结构清晰；
• Output quality can be measured objectively.
  输出质量可以客观衡量。

In our own implementation, agents can now solve real GitHub issues in the SWE-bench Verified benchmark based on the pull request description alone. However, whereas automated testing helps verify functionality, human review remains crucial for ensuring solutions align with broader system requirements.
在我们的实现中，代理现在能够仅根据拉取请求描述解决 SWE-bench Verified 基准中的真实 GitHub 问题。然而，尽管自动化测试有助于验证功能，但人工审查对于确保解决方案符合更广泛的系统需求仍然至关重要。

Appendix 2: Prompt engineering your tools
附录 2：工具的提示工程

No matter which agentic system you're building, tools will likely be an important part of your agent. Tools enable Claude to interact with external services and APIs by specifying their exact structure and definition in our API. When Claude responds, it will include a tool use block in the API response if it plans to invoke a tool. Tool definitions and specifications should be given just as much prompt engineering attention as your overall prompts. In this brief appendix, we describe how to prompt engineer your tools.
无论你构建的是哪种代理系统，工具都可能成为你代理的重要组成部分。工具使 Claude 能够通过在我们的 API 中精确指定其结构和定义，与外部服务和 API 进行交互。当 Claude 响应时，如果它计划调用工具，API 响应中将包含一个工具使用块。工具的定义和规范应与整体提示一样，给予同等程度的提示工程关注。在这份简短的附录中，我们将描述如何对工具进行提示工程设计。

There are often several ways to specify the same action. For instance, you can specify a file edit by writing a diff, or by rewriting the entire file. For structured output, you can return code inside markdown or inside JSON. In software engineering, differences like these are cosmetic and can be converted losslessly from one to the other. However, some formats are much more difficult for an LLM to write than others. Writing a diff requires knowing how many lines are changing in the chunk header before the new code is written. Writing code inside JSON (compared to markdown) requires extra escaping of newlines and quotes.
指定同一操作通常有多种方式。例如，你可以通过编写 diff 来指定文件编辑，或者通过重写整个文件来实现。对于结构化输出，你可以在 markdown 中嵌入代码，也可以在 JSON 中嵌入代码。在软件工程中，这类差异大多是表面性的，可以无损地从一种形式转换到另一种形式。然而，某些格式对LLM来说比其他格式更难编写。编写 diff 需要在新代码编写之前知道块头中更改的行数。在 JSON 中编写代码（与 markdown 相比）需要额外对换行符和引号进行转义。

Our suggestions for deciding on tool formats are the following:
我们建议在选择工具格式时考虑以下几点：

• Give the model enough tokens to "think" before it writes itself into a corner.
  给模型足够的令牌来“思考”，以免它把自己写进死胡同。
• Keep the format close to what the model has seen naturally occurring in text on the internet.
  保持格式接近于模型在互联网文本中自然出现的形式。
• Make sure there's no formatting "overhead" such as having to keep an accurate count of thousands of lines of code, or string-escaping any code it writes.
  确保没有诸如需要精确统计成千上万行代码或对其所写代码进行字符串转义之类的“额外开销”。

One rule of thumb is to think about how much effort goes into human-computer interfaces (HCI), and plan to invest just as much effort in creating good agent-computer interfaces (ACI). Here are some thoughts on how to do so:
一个经验法则是考虑在人机交互界面（HCI）上投入了多少精力，并计划在创建良好的代理计算机界面（ACI）上投入同样多的精力。以下是一些关于如何做到这一点的思考：

• Put yourself in the model's shoes. Is it obvious how to use this tool, based on the description and parameters, or would you need to think carefully about it? If so, then it’s probably also true for the model. A good tool definition often includes example usage, edge cases, input format requirements, and clear boundaries from other tools.
  设身处地为模型着想。根据描述和参数，使用这个工具是否显而易见，还是需要仔细思考？如果是后者，那么对模型来说可能也是如此。一个好的工具定义通常包括示例用法、边缘情况、输入格式要求以及与其他工具的明确界限。
• How can you change parameter names or descriptions to make things more obvious? Think of this as writing a great docstring for a junior developer on your team. This is especially important when using many similar tools.
  如何更改参数名称或描述以使内容更加清晰？将此视为为团队中的初级开发人员编写出色的文档字符串。在使用众多相似工具时，这一点尤为重要。
• Test how the model uses your tools: Run many example inputs in our workbench to see what mistakes the model makes, and iterate.
  测试模型如何使用您的工具：在我们的工作台中运行多个示例输入，以观察模型所犯的错误，并进行迭代改进。
• Poka-yoke your tools. Change the arguments so that it is harder to make mistakes.
  将你的工具防错化。调整参数，以减少出错的可能性。

While building our agent for SWE-bench, we actually spent more time optimizing our tools than the overall prompt. For example, we found that the model would make mistakes with tools using relative filepaths after the agent had moved out of the root directory. To fix this, we changed the tool to always require absolute filepaths—and we found that the model used this method flawlessly.
在构建 SWE-bench 的代理时，我们实际上花费了更多时间优化工具，而非整体提示。例如，我们发现模型在使用相对文件路径时会出现错误，尤其是在代理移出根目录后。为解决此问题，我们将工具修改为始终要求绝对文件路径，结果发现模型能完美运用此方法。