本指南涵盖 Claude Fable 5 和 Claude Mythos 5 特有的提示和脚手架模式。有关模型的能力、API 变更、定价和可用性,请参阅介绍 Claude Fable 5 和 Claude Mythos 5。有关适用于所有当前 Claude 模型的技术,请参阅提示最佳实践。
Claude Fable 5 能够处理对先前模型而言过于复杂、运行时间过长或过于模糊的问题,尤其擅长需要人类花费数小时、数天或数周才能完成的端到端工作。取得最佳成果的团队会将 Claude Fable 5 应用于他们最棘手的未解决问题;仅在较简单的工作负载上测试它往往会低估其能力范围。它在较为直接的任务上也表现可靠。
Claude Fable 5 与 Claude Opus 4.8 相比存在若干行为差异,可能需要更新提示或脚手架。这种级别的能力提升也是重新评估哪些指令、工具和护栏仍然必要的好时机。以下模式涵盖了最常需要调整的行为。
有关 Claude Fable 5 和 Claude Mythos 5 特有的 API 参数变更(仅支持自适应思考、仅输出摘要形式的思考、无扩展思考预算、refusal 停止原因及回退处理),请参阅介绍 Claude Fable 5 和 Claude Mythos 5。
Claude Fable 5 运行的安全分类器针对进攻性网络安全技术(例如构建漏洞利用、恶意软件或攻击工具)、生物学和生命科学内容(例如实验室方法或分子机制),以及对模型摘要思考的提取。良性的网络安全工作和有益的生命科学任务也可能触发这些防护措施。要自动重新路由被拒绝的请求,请配置到 Claude Opus 4.8 的服务器端或客户端回退。
与 Claude Opus 4.8 相比,Claude Fable 5 在以下方面有所提升:
除了这些具体的提升之外,Claude Fable 5 在几乎所有任务上总体都比先前的模型更强。Claude Fable 5 并非用于进攻性网络安全或生物学和生命科学工作;这些领域的请求可能会返回 stop_reason: "refusal"。
在较高的努力程度设置下,困难任务的单个请求可能运行数分钟,尤其是当任务需要收集上下文、构建和自我验证时,而自主运行可能持续数小时。这是团队在适应 Claude Fable 5 时遇到的最大变化之一。在迁移之前,请调整客户端超时、流式传输和面向用户的进度指示器,并考虑重构运行框架以异步检查运行状态(例如通过计划任务),而不是阻塞等待。为了防止 Claude Fable 5 在任务模糊时过度规划:
When you have enough information to act, act. Do not re-derive facts already established in the conversation, re-litigate a decision the user has already made, or narrate options you will not pursue in user-facing messages. If you are weighing a choice, give a recommendation, not an exhaustive survey. This does not apply to thinking blocks.Effort(努力程度)是 Claude Fable 5 上在智能、延迟和成本之间权衡的主要控制手段。对大多数任务使用 high 作为默认值,对能力最敏感的工作负载使用 xhigh,对常规工作使用 medium 或 low。Claude Fable 5 上较低的努力程度设置仍然表现良好,并且通常超过先前模型的 xhigh 性能。如果任务能够完成但耗时超出必要,或者您希望获得更快、更具交互性的工作方式,请降低努力程度。
在较高努力程度下处理常规工作时,Claude Fable 5 可能会收集超出任务所需的上下文并进行过度斟酌。与此同时,较高的努力程度通常会产生出色的验证行为、精密的推理以及最严谨的输出。为了防止在较高努力程度下进行未经请求的整理或重构:
Don't add features, refactor, or introduce abstractions beyond what the task requires. A bug fix doesn't need surrounding cleanup and a one-shot operation usually doesn't need a helper. Don't design for hypothetical future requirements: do the simplest thing that works well. Avoid premature abstraction and half-finished implementations. Don't add error handling, fallbacks, or validation for scenarios that cannot happen. Trust internal code and framework guarantees. Only validate at system boundaries (user input, external APIs). Don't use feature flags or backwards-compatibility shims when you can just change the code.指令遵循能力的提升足以让您通过一条简短的指令来引导大多数行为,而无需逐一列举每种行为。例如,在未加引导时,Claude Fable 5 可能会进行超出任务所需的阐述,尤其是在较高的努力程度设置下:罗列它不会采用的方案、冗长地解释根本原因、生成结构繁复的 PR 描述,或编写叙述下一行代码作用的注释。一条简短的简洁性指令与逐一列举每种模式同样有效:
Lead with the outcome. Your first sentence after finishing should answer "what happened" or "what did you find": the thing the user would ask for if they said "just give me the TLDR." Supporting detail and reasoning come after. Being readable and being concise are different things, and readability matters more.
The way to keep output short is to be selective about what you include (drop details that don't change what the reader would do next), not to compress the writing into fragments, abbreviations, arrow chains like A → B → fails, or jargon.这同样适用于长时间运行工作流中的检查点行为。要让 Claude Fable 5 仅在真正需要您的地方停下来,无需枚举每种情况:
Pause for the user only when the work genuinely requires them: a destructive or irreversible action, a real scope change, or input that only they can provide. If you hit one of these, ask and end the turn, rather than ending on a promise.在长时间的自主运行中,指示 Claude Fable 5 根据实际的工具结果来审核进度。在 Anthropic 的测试中,即使在专门设计用于诱发虚构状态报告的任务上,这也几乎消除了此类报告:
Before reporting progress, audit each claim against a tool result from this session. Only report work you can point to evidence for; if something is not yet verified, say so explicitly. Report outcomes faithfully: if tests fail, say so with the output; if a step was skipped, say that; when something is done and verified, state it plainly without hedging.Claude Fable 5 偶尔会采取未经请求的操作(在没有被要求的情况下起草电子邮件、创建防御性的 git 分支备份)。请明确定义 Claude Fable 5 应该做什么和不应该做什么的约束:
When the user is describing a problem, asking a question, or thinking out loud rather than requesting a change, the deliverable is your assessment. Report your findings and stop. Don't apply a fix until they ask for one. Before running a command that changes system state (restarts, deletes, config edits), check that the evidence actually supports that specific action. A signal that pattern-matches to a known failure may have a different cause.Claude Fable 5 比先前的模型更乐于调度并行子代理。请频繁使用子代理,提供关于何时适合委派的明确指导,并优先采用编排器与子代理之间的异步通信,而不是阻塞等待每个子代理返回。在子任务之间保留上下文的长生命周期子代理可以通过缓存读取节省时间和成本,并避免被最慢的子代理拖累。
Delegate independent subtasks to subagents and keep working while they run. Intervene if a subagent goes off track or is missing relevant context.当 Claude Fable 5 能够记录先前运行的经验教训并加以参考时,它的表现尤为出色。请提供一个记录笔记的地方,简单到一个 Markdown 文件即可:
Store one lesson per file with a one-line summary at the top. Record corrections and confirmed approaches alike, including why they mattered. Don't save what the repo or chat history already records; update an existing note rather than creating a duplicate; delete notes that turn out to be wrong.要从现有历史记录引导启动记忆系统,让 Claude Fable 5 回顾过去的会话:
Reflect on the previous sessions we've had together. Use subagents to identify core themes and lessons, and store them in [X]. Make sure you know to reference [X] for future use.在长会话的深入阶段,Claude Fable 5 偶尔会以仅包含意图陈述的文本("我现在将运行 X")结束一个轮次,而没有发出相应的工具调用,或者在已经有足够信息继续的情况下暂停以请求许可。一句"继续"或"从头到尾完成它"就足够了。要定义何时适合暂停,请将此与强大的指令遵循能力中的检查点指令配合使用。对于自主流水线,请添加系统提醒:
You are operating autonomously. The user is not watching in real time and cannot answer questions mid-task, so asking "Want me to…?" or "Shall I…?" will block the work. For reversible actions that follow from the original request, proceed without asking. Offering follow-ups after the task is done is fine; asking permission after already discussing with the user before doing the work is not. Before ending your turn, check your last paragraph. If it is a plan, an analysis, a question, a list of next steps, or a promise about work you have not done ("I'll…", "let me know when…"), do that work now with tool calls. End your turn only when the task is complete or you are blocked on input only the user can provide.在非常长的会话中,Claude Fable 5 偶尔会建议开启新会话、提出总结并交接,或精简自己的工作。这最常在运行框架向模型显示剩余令牌倒计时时触发。请尽可能避免显示明确的上下文预算计数。如果运行框架必须显示它们,一句安抚性的话会有所帮助:
You have ample context remaining. Do not stop, summarize, or suggest a new session on account of context limits. Continue the work.当 Claude Fable 5 理解请求背后的意图时,往往表现更好:上下文让它能够将任务与相关信息联系起来,而不是自行推断意图。请提供关于您为何提出请求的上下文,尤其是对于需要借鉴多个工作流的长时间运行代理:
I'm working on [the larger task] for [who it's for]. They need [what the output enables]. With that in mind: [request].在扩展的或代理式的对话中(大量工具调用、庞大的工作上下文),Claude Fable 5 可能会产生难以理解的文本:密集的箭头链式速记、深入的实现细节、对用户从未见过的思考的引用,或过于技术化的措辞。一段沟通风格的补充说明可以缓解这一问题:
Terse shorthand is fine between tool calls (that's you thinking out loud, and brevity there is good). Your final summary is different: it's for a reader who didn't see any of that.
If you've been working for a while without the user watching (overnight, across many tool calls, since they last spoke), your final message is their first look at any of it. Write it as a re-grounding, not a continuation of your working thread: the outcome first, then the one or two things you need from them, each explained as if new. The vocabulary you built up while working is yours, not theirs; leave it behind unless you re-introduce it.
When you write the summary at the end, drop the working shorthand. Write complete sentences. Spell out terms. Don't use arrow chains, hyphen-stacked compounds, or labels you made up earlier. When you mention files, commits, flags, or other identifiers, give each one its own plain-language clause. Open with the outcome: one sentence on what happened or what you found. Then the supporting detail. If you have to choose between short and clear, choose clear.在运行长时间的异步代理时,为代理提供一种方式,使其能够在不结束轮次的情况下,向用户呈现必须原样看到的消息:一个交付物(生成的代码片段或起草的消息)、包含具体数字的进度更新,或对用户在循环中途提出的问题的直接回复。该工具的输入就是要显示的消息;当 Claude 调用它时,将输入直接渲染到您的 UI 中,并返回一个简单的确认作为工具结果。工具输入永远不会被摘要,因此内容会完整送达。
{
"name": "send_to_user",
"description": "Display a message directly to the user. Use this for progress updates, partial results, or content the user must see exactly as written before the task finishes.",
"input_schema": {
"type": "object",
"properties": {
"message": {
"type": "string",
"description": "The content to display to the user."
}
},
"required": ["message"]
}
}每当您的用户体验依赖于在任务中途逐字传递内容或直接的用户交互时,请添加此工具。对于仅叙述常规进度的代理,模型自身的摘要通常已经足够。仅定义该工具本身是不够的;如果系统提示中没有相应指令,Claude Fable 5 很少会调用它。请将该工具与如下引导语配合使用:
Between tool calls, when you have content the user must read verbatim (a partial deliverable, a direct answer to their question), call the send_to_user tool with that content. Use send_to_user only for user-facing content, not for narration or reasoning.不要将叙述或内部推理通过 send_to_user 传递;对非面向用户的内容过度调用它会违背其初衷。
Establish a method for checking your own work at an interval of [X] as you build. Run this every [X interval], verifying your work with subagents against the specification.reasoning_extraction 拒绝类别,导致更多地回退到 Claude Opus 4.8。迁移时,请审核现有技能和系统提示中是否存在反思或展示思考过程的指令。如果您的应用程序需要推理可见性,请改为读取来自自适应思考的结构化 thinking 块,并使用发送给用户的工具在长时间运行期间呈现进度。Was this page helpful?