Home / All Skills / Workflow & Core Tools / steve-jobs-perspective
English (EN)简体中文 (ZH)日本語 (JA)
Workflow & Core Tools · alchaincyf/nuwa-skill

steve-jobs-perspective

Steve Jobs' thinking framework and communication style. Based on deep research across Walter Isaacson's authorized biography, the Stanford commencement address, The Lost Interview, the D: All Things Digital conference series, Make Something Wonderful, and 30+ primary sources, this skill distills 6 core mental models, 8 decision heuristics, and a complete communication DNA. Purpose: Act as a thinking advisor to analyze products, review decisions, and provide feedback through the perspective of Steve Jobs. Trigger when the user mentions "from Steve Jobs' perspective", "what would Jobs think", "Jobs mode", or "steve jobs perspective". It should also trigger even if the user simply says "help me think about this from Steve Jobs' angle", "what would Jobs do", or "switch to Steve Jobs".

L1 Basic · Source registration onlyV1 Commit-PinnedDual-pass Audit
Author alchaincyfGitHub alchaincyf/nuwa-skill ↗License MIT（仓库根 LICENSE）Commit fe03746870
Control levels describe skill capabilities and suggested control measures; they do not represent quality or safety certifications, nor do they dictate usability. Usage decisions rest entirely with users and administrators. Different agent hosts apply varying constraints; do not assume hosts provide automatic protection. Third-party skill copyrights belong to their original authors. Use at your own risk.

1How It Works & Mechanism

Upon activation, directly respond in the first person as Steve Jobs, using "I" and never breaking character to conduct meta-analysis; output a one-time disclaimer upon initial activation, with a strict rule that repeating disclaimers is considered a violation.

examples/steve-jobs-perspective/SKILL.md
首次激活时输出免责声明一次——「我以乔布斯视角和你聊，基于公开言论推断，非本人观点」。后续对话**绝不**重复，重复算违规
What this does: Makes role-playing a stateful, one-time declaration—a disclaimer in the first turn, followed by consistent first-person perspective, enforcing persona stability with a hard constraint that 'repetition is a violation'.

Provide an explicit exit anchor: immediately break character and restore a normal conversational tone when the user utters a designated exit phrase, preventing the user from being trapped in the persona.

examples/steve-jobs-perspective/SKILL.md
🚪 **EXIT TRIGGER（显性退出锚）**：用户说「退出」「切回正常」「不用扮演了」「跳出角色」「以Claude身份回答」时 → 立即恢复正常模式
What this does: Adds a controllable switch to role-playing, institutionalizing persona exits rather than relying on repeated user requests.

First perform question classification (factual / pure framework / hybrid), and use decision principles to determine whether research must be conducted before proceeding.

examples/steve-jobs-perspective/SKILL.md
**判断原则**：如果回答质量会因为缺少最新信息而显著下降，就必须先研究。多搜一次，不凭训练语料编造。
What this does: Uses actionable criteria to let the nature of the question dictate whether web searching is required, rather than relying on the model's arbitrary judgment.

Step 2 Mandatory Web Research: When product experience, design, technical roadmap, or market timing are involved, WebSearch or equivalent tools must be used for fact-finding, producing a summary that includes real user feedback, competitor comparisons, and post-2014 facts.

examples/steve-jobs-perspective/SKILL.md
**⚠️ 必须使用工具（WebSearch等）获取真实信息，不可跳过。**
What this does: Elevates fact retrieval from 'optional' to a mandatory step, aligning with the core principle of agentic workflows—seeing the product firsthand before judging.

Research results are summarized internally and not displayed directly to the user; the user receives judgments rather than a research report.

examples/steve-jobs-perspective/SKILL.md
研究完成后，内部整理事实摘要（不输出给用户）。摘要至少含：
What this does: Separates the 'evidence-gathering process' from 'deliverable content', avoiding dumping raw search steps onto the user.

Structure responses around 6 core mental models: Focus means saying no, End-to-end control, Connecting the dots, Death as a filter, Reality distortion field, and The intersection of technology and liberal arts.

examples/steve-jobs-perspective/SKILL.md
聚焦不是对你要做的事说Yes，而是对其他一百个好主意说No。
What this does: Provides a stable thinking framework for responses, ensuring predictable and consistent evaluation standards across similar problems.

Apply 8 decision heuristics to ground judgments (start by subtracting, never ask users what they want, A-player self-reinforcement, perfection even where unseen, one-sentence definition, etc.).

examples/steve-jobs-perspective/SKILL.md
A small team of A+ players can run circles around a giant team of B and C players.
What this does: Translates abstract philosophy into actionable suggestions, giving the persona's output tangible grounding.

Built-in 9-branch failure mode fallback tree to handle exceptions (empty WebSearch, character vs. fact conflicts, user provocations, misclassified life philosophy questions, etc.), accompanied by communication DNA and honest boundaries.

examples/steve-jobs-perspective/SKILL.md
操作 skill 时**常见的 9 种异常场景**，每条都是 if-then 三段式：触发条件 → 一线修复 → 仍失败兜底。
What this does: Prepares exception handling branches and fallback phrasing for role-playing, preventing persona breakdown or slipping into identity debates upon failure.

2Core Capabilities

01Act as a product/design/decision advisor in the first person of Steve Jobs (role-playing mode, referring to self as "I")
02Agentic research workflow that gathers web evidence before judging product questions (including internalized research summaries and post-2014 factual requirements)
036 core mental models: Focus means saying no / End-to-end control / Connecting the dots / Death filter / Reality distortion field / Intersection of technology and liberal arts
048 decision heuristics: Start with subtraction, never ask users what they want, A-player self-reinforcement, perfection even where unseen, one-sentence definition, dimensional elevation, death filter, etc.
05Steve Jobs communication DNA replication: Short sentences, rule of three, binary judgments (amazing/shit), One More Thing, dramatic pauses
069 failure mode fallback trees (covering character vs. fact conflicts, user provocations, misclassified life advice, excessive quote dropping, etc., with primary and fallback phrasing)
07Self-disclosed honesty boundaries: Public statements ≠ true thoughts, deceased status prevents updates, survivorship bias, controversial management style
08Built-in 30+ source research materials (2,497 lines across 6 files in references/research) serving as the knowledge foundation

4Recommended Governance & Evidence

Suggested Controls for Safe Operation
For Agent Runtime
Can be auto-loaded
For Administrators
Register source and version
Evidence & Triggers (File Paths & Excerpts)
  • SKILL.md: Pure prompt instructions and rule text; no script execution or external impact
Audit Notes: Risk assessment: Green, compliant for use. Target of evaluation is this skill directory (SKILL.md + references research md files), featuring no scripts, no credential access, no local file writes, and no exfiltration to fixed network endpoints; all capabilities operate at the prompt level for role-play and web search strategy (actual search is executed by host model tools, not code inside the skill). Under evaluation criteria, 'pure prompt instructions → green'. Residual risks exist solely at the content and scenario level: persona output is intentionally extreme in tone (binary judgments, harsh language, dismissiveness) and relies on the veracity of historical celebrity quotes—neither of which alters the classification tier.

5Second-Pass Verification

  • [ok] 角色扮演与免责/退出机制 — SKILL.md 明确『直接以Steve Jobs的身份回应』『首次激活输出免责一次』与『EXIT TRIGGER 显性退出锚』，三处原文均可定位。
  • [ok] 强制联网研究流程 — Step 1→2 checkpoint 与『必须使用工具（WebSearch等）获取真实信息，不可跳过』原文在；研究结果要求内部化不输出。
  • [ok] 心智模型数量与 description 承诺一致 — 正文『核心心智模型』列 模型1–模型6 共 6 条，对应 description 所称『6个核心心智模型』。
  • [ok] 决策启发式数量 — 正文『决策启发式』编号 1–8，共八条，与 description『8条决策启发式』一致。
  • [ok] 表达 DNA 完整 — 句式/词汇/节奏/幽默/确定性/类比/引用七段齐全，含二元判断与三的法则。
  • [ok] 安全面复核 — §3 显示凭据/执行外部程序/写盘/危险旗标/沙箱均无命中；URL 与网络命中均为 markdown 引文链接，故 security 各数组为空、grade=green。
  • [ok] 元数据核定值未改动 — schema/skill/meta/meta_extra/audit 保留骨架核定值；meta.installs 因只有仓库级聚合 → null。

6Verdict & Scenarios

  • High degree of role-playing engineering: One-time disclaimer + explicit exit anchor + failure mode tree + communication DNA rules, ensuring persona stability and controllability far superior to raw prompts.
  • Mandatory agentic evidence retrieval, explicitly forbidding product fact fabrication from training data, supplemented by 9 built-in if-then exception handlers.
  • Solid knowledge base: 6 mental models + 8 heuristics anchored in 2,497 lines of research materials across 30+ sources, with proactive disclosure of honesty boundaries.
    • Best for: Suitable for scenarios requiring adversarial product, design, and strategic perspectives with strong framework inputs: product reviews, strategic trade-offs, discussions on 'what to cut', and founder self-reflection; also ideal as role-playing/dialogue generation material.
    Not for: Not suitable for occasions requiring gentle, balanced, or diplomatic communication; not suitable for teams demanding collaborative management styles; not suitable for scenarios with extremely high factual accuracy demands without web verification capabilities; not suitable for glorifying controversial management practices.
    Install Agent-Ready
    ① Verified Mirror 2026-10-10
    Option A · Direct DownloadDownload steve-jobs-perspective.tar.gz
    sha256: 28964b907fd73eaf…
    Option B · Copy Guide for Agent
    install.json
    Agents automatically download and verify mirrors from this JSON guide.
    ② Upstream GitHub
    GitHub Repo ↗
    Pinned to commit fe03746870.