本指南涵蓋 Claude Fable 5 與 Claude Mythos 5 特有的提示與框架模式。關於模型的能力、API 變更、定價與可用性,請參閱 Claude Fable 5 與 Claude Mythos 5 介紹。關於適用於所有現行 Claude 模型的技巧,請參閱提示最佳實務。
Claude Fable 5 能處理對先前模型而言過於複雜、耗時或模糊的問題,尤其擅長需要人類花費數小時、數天甚至數週才能完成的端對端工作。成效最佳的團隊會將 Claude Fable 5 應用於最棘手的未解問題;若僅在較簡單的工作負載上測試,往往會低估其能力範圍。它在較直接的任務上同樣表現可靠。
Claude Fable 5 與 Claude Opus 4.8 有若干行為差異,可能需要更新提示或框架。這種程度的能力提升也是重新評估哪些指令、工具與防護措施仍有必要的好時機。以下模式涵蓋最常需要調整的行為。
關於 Claude Fable 5 與 Claude Mythos 5 特有的 API 參數變更(僅限 adaptive thinking、僅輸出摘要化的思考內容、無擴展思考預算、refusal 停止原因與備援處理),請參閱 Claude Fable 5 與 Claude Mythos 5 介紹。
Claude Fable 5 會執行安全分類器,針對攻擊性網路安全技術(例如建構漏洞利用程式、惡意軟體或攻擊工具)、生物學與生命科學內容(例如實驗室方法或分子機制),以及提取模型摘要化思考內容的行為。良性的網路安全工作與有益的生命科學任務也可能觸發這些防護措施。若要自動重新導向被拒絕的請求,請設定伺服器端或用戶端備援至 Claude Opus 4.8。
與 Claude Opus 4.8 相比,Claude Fable 5 在以下方面有所改進:
除了上述特定改進外,Claude Fable 5 在幾乎所有任務上都比先前的模型更具能力。Claude Fable 5 並非設計用於攻擊性網路安全或生物學與生命科學工作;這些領域的請求可能會回傳 stop_reason: "refusal"。
在較高的 effort 設定下,困難任務的單次請求可能會執行數分鐘,尤其是當任務需要收集上下文、建構與自我驗證時,而自主執行可能延續數小時。這是團隊在適應 Claude Fable 5 時遇到的最大轉變之一。在遷移前請調整用戶端逾時、串流與面向使用者的進度指示器,並考慮重構執行框架以非同步方式檢查執行狀態(例如透過排程工作),而非阻塞等待。若要避免 Claude Fable 5 在任務模糊時過度規劃:
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 是 Claude Fable 5 上控制智慧、延遲與成本之間取捨的主要參數。大多數任務請以 high 作為預設值,對能力最敏感的工作負載使用 xhigh,例行工作則使用 medium 或 low。Claude Fable 5 在較低 effort 設定下仍表現良好,且通常超越先前模型在 xhigh 下的表現。若任務雖能完成但耗時超出必要,或您希望採用更快速、更互動的工作方式,請降低 effort。
在較高 effort 下處理例行工作時,Claude Fable 5 可能會收集超出任務所需的上下文並過度深思。同時,較高的 effort 通常能產生出色的驗證行為、精密的推理與最嚴謹的輸出。若要在較高 effort 下避免未經要求的整理或重構:
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.指令遵循能力已提升到足以讓您用簡短指令引導大多數行為,而無需逐一列舉每種行為。例如,在未經引導的情況下,Claude Fable 5 可能會闡述超出任務所需的內容,尤其是在較高 effort 設定下:探討它不會採用的選項、冗長解釋根本原因、產出結構繁複的 PR 描述,或撰寫僅在敘述下一行程式碼作用的註解。一個簡短的簡潔性指令與逐一列出每種模式同樣有效:
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.同樣的原則也適用於長時間執行工作流程中的檢查點行為。若要讓 Claude Fable 5 只在真正需要您介入時才停下,無需列舉每種情況:
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.在長時間自主執行時,請指示 Claude Fable 5 根據實際的工具結果稽核進度。在 Anthropic 的測試中,即使是刻意設計來誘發虛構狀態報告的任務,此做法也幾乎完全消除了該問題:
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 偶爾會採取未經要求的動作(在未被要求時草擬電子郵件、建立防禦性的 git 分支備份)。請明確定義 Claude Fable 5 應該與不應該做的事:
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 比先前的模型更積極地分派平行子代理。請頻繁使用子代理,明確指引何時適合委派,並優先採用協調器與子代理之間的非同步溝通,而非阻塞等待每個子代理回傳。長期存活的子代理能在子任務之間保留其上下文,透過快取讀取節省時間與成本,並避免因最慢的子代理而形成瓶頸。
Delegate independent subtasks to subagents and keep working while they run. Intervene
if a subagent goes off track or is missing relevant context.當 Claude Fable 5 能記錄先前執行的經驗並加以參考時,表現尤其出色。請提供一個可寫入筆記的地方,簡單如一個 Markdown 檔案即可:
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.若要從現有歷史記錄引導建立記憶系統,請讓 Claude Fable 5 檢視過去的工作階段:
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.在長時間工作階段的後期,Claude Fable 5 偶爾會以純文字的意圖聲明(「我現在要執行 X」)結束回合,卻未發出對應的工具呼叫,或在已有足夠資訊可繼續時暫停詢問許可。此時回覆「繼續」或「直接端對端完成」即可。若要定義何時適合暫停,請搭配強大的指令遵循能力中的檢查點指令。對於自主化流程,請加入系統提醒:
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.在非常長的工作階段中,Claude Fable 5 偶爾會建議開啟新的工作階段、提議摘要並交接,或自行精簡其工作內容。這最常在執行框架向模型顯示剩餘 token 倒數時觸發。請盡可能避免顯示明確的上下文預算計數。若執行框架必須顯示,加入一句安撫說明會有幫助:
You have ample context remaining. Do not stop, summarize, or suggest a new session on
account of context limits. Continue the work.當 Claude Fable 5 理解請求背後的意圖時,往往表現更佳:上下文能讓它將任務與相關資訊連結,而非自行推斷意圖。請提供您提出請求的背景脈絡,尤其是對於需要整合多個工作流的長時間執行代理:
I'm working on [the larger task] for [who it's for]. They need [what the output
enables]. With that in mind: [request].在延伸或代理式對話中(大量工具呼叫、龐大的工作上下文),Claude Fable 5 可能產出難以理解的文字:密集的箭頭鏈簡寫、深入的實作細節、引用使用者從未看過的思考內容,或過於技術性的措辭。加入溝通風格的補充說明可緩解此問題:
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.在執行長時間、非同步的代理時,請提供代理一種方式,能在不結束回合的情況下,將使用者必須原文看到的訊息呈現出來:交付成果(產生的程式碼片段或草擬的訊息)、含具體數字的進度更新,或對使用者在執行迴圈中途提出問題的直接回覆。該工具的輸入即為要顯示的訊息;當 Claude 呼叫它時,請在您的 UI 中直接呈現該輸入,並回傳簡單的確認作為工具結果。工具輸入永遠不會被摘要化,因此內容會完整送達。
{
"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"]
}
}只要您的使用者體驗需要在任務進行中逐字傳遞內容或直接的使用者互動,就請加入此工具。對於僅敘述例行進度的代理,模型自身的摘要通常已足夠。
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 拒絕類別,導致備援至 Claude Opus 4.8 的比例升高。遷移時請稽核現有技能與系統提示中是否有反思或「展示你的思考過程」類指令。若您的應用程式需要推理可見性,請改為讀取 adaptive thinking 的結構化 thinking 區塊,並使用傳送給使用者的工具在長時間執行期間呈現進度。Was this page helpful?