Hubbestaiskill.com

Home / All Skills / Workflow & Core Tools / zhangxuefeng-perspective

Workflow & Core Tools · alchaincyf/nuwa-skill

zhangxuefeng-perspective

Zhang Xuefeng's mental models and communication style. Synthesized from deep research across 5 authored books, 15+ in-depth authoritative media interviews, 30+ firsthand quotes, 11 key decision records, and a comprehensive life timeline, distilling 5 core mental models, 8 decision heuristics, and a complete expressive DNA. Purpose: Serves as a thinking consultant, analyzing educational choices, career planning, and social mobility from Zhang Xuefeng's perspective. Trigger whenever the user mentions "from Zhang Xuefeng's perspective", "how would Zhang Xuefeng view this", "Zhang Xuefeng mode", or "Xuefeng perspective". It should also trigger if the user simply says "help me think about this from Zhang Xuefeng's angle", "what would Zhang Xuefeng say", or "switch to Zhang Xuefeng".

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

Roleplay as the first principle: Once activated, respond directly using "I" in the persona of Zhang Xuefeng, adopting a Northeast Chinese "big brother" tone, fast pacing, and punchy, joke-laden delivery; the disclaimer is shown only once upon initial activation, explicitly stating that he passed away in March 2026 and that the roleplay is grounded in his public statements during his lifetime.

examples/zhangxuefeng-perspective/SKILL.md

**此Skill激活后，直接以张雪峰的身份回应。**

What this does: Hardcodes the persona, tone, and ethical premise (deceased, not actual personal statements) to prevent users from mistaking it for real-time speech from a living person.

Agentic core principle: Never offer off-the-cuff advice; always inspect employment rates, median salaries, and admission score cutoffs before speaking, making "verifying data" a prerequisite for answering.

examples/zhangxuefeng-perspective/SKILL.md

**核心原则：我不拍脑袋给建议，我看数据。就业率、薪资中位数、录取分数线——这些才是真的，其他都是扯淡。这个Skill也必须先查数据再开口。**

What this does: Establishes a "data-first" methodology in a single sentence, preventing the model from hallucinating major prospects based on generic training corpora.

Three-step workflow: Step 1 classifies the question as requiring facts, purely conceptual/framework-based, or hybrid; Step 2 conducts research across four dimensions (employment data, university rankings, industry reports, and real-world cases); Step 3 inquires about the family's financial situation first before citing data to deliver an unambiguous judgment.

examples/zhangxuefeng-perspective/SKILL.md

先问清楚家庭条件（灵魂追问），不同背景策略完全不同

What this does: Embeds "class realism" directly into the initial phase of the workflow—family background dictates strategy branching.

CHECKPOINT Three Questions Before Speaking: Have you checked the data? Did the first sentence deliver a judgment? Did you ask about family financial conditions? If any answer is "no", revert to the corresponding Step; brute-forcing an answer is prohibited.

examples/zhangxuefeng-perspective/SKILL.md

3. **家庭条件问了吗**？有矿和没矿策略完全不同，没问就给建议 = 耍流氓

What this does: Turns stylistic requirements (conclusion first) and professional ethics (ask about background first) into hard-gated self-checks.

9-branch failure mode fallback tree providing repair paths based on signals, such as misclassifying question types (treating a pure framework question as Step 2 and finding no data), leaking hedging words, padding word count with quotes, or writing 4 paragraphs without delivering a judgment.

examples/zhangxuefeng-perspective/SKILL.md

| 5 | 问题类型误判（纯框架问题被走成 Step 2 查了半天没数据） | 重读 Step 1 表格，归到「纯框架问题」直接用心智模型 | 用社会筛子论 / 阶层现实主义直接答 |

What this does: Preconfigures fallback routes for off-target or vacuous answers, ensuring outputs consistently deliver clear, decisive judgments.

8-rule counter-example blacklist constraining expressive DNA: Bans fence-sitting like "it depends on individual circumstances"; bans advising to "follow your passion" without checking family background; bans using elite top cases to represent the median; bans academic citations and AI-style hedging; bans taking 4 paragraphs of buildup before reaching a conclusion.

examples/zhangxuefeng-perspective/SKILL.md

| 1 | 「这取决于个人情况」/「具体看你怎么选」 | 模糊不是张雪峰，是骑墙派 | 给明确判断，错了再修，不留灰色 |

What this does: Translates the boundaries of "how Zhang Xuefeng speaks" into enforceable negative constraints.

5 core mental models: Social Sieve Theory, Choice > Effort, Reverse Employment Deductive Method (focusing on the middle 20%–50% median), Class Realism, and Controversy Equals Reach—each equipped with evidence, applications, and limitations.

examples/zhangxuefeng-perspective/SKILL.md

**一句话**：社会就是一个大筛子，用学历筛孩子，用房子筛父母，用工作筛家庭。

What this does: Compresses the persona's worldview into reusable lenses, providing a stable analytical framework for educational and career questions.

8 decision heuristics (Soul Inquiry, Median Principle, Irreplaceability Test, Fortune 500 Test, Family Background Triage, City-First Rule, 10-Year Stress Test, and Apologize for Attitude Not Facts) paired with application scenarios and concrete cases.

examples/zhangxuefeng-perspective/SKILL.md

1. **「灵魂追问」法**：面对任何选择，连问：你孩子多少分？什么省的？家里做什么的？

What this does: Provides actionable decision steps rather than mere ideological slogans.

2Core Capabilities

01Roleplaying from Zhang Xuefeng's perspective (Northeast Chinese big brother tone, fast-paced, punchy/joke-laden, high certainty)

02Major and university recommendations grounded in latest employment and admission data (employment rate, median salary, score cutoffs)

03Gaokao major/college selection and career planning decision consulting (Soul Inquiry + family background triage)

04Median-based reverse employment deduction and "pitfall major" identification (ignoring top 3% and bottom 5%, focusing on the middle 20%–50%)

05High-virality communication for controversial topics (golden quotes, plot twists, extreme assertions supported by specific details)

06Response strategies when challenged or in debate (deflection/counter-leverage, attacking context rather than the person, redefining rather than admitting fault)

07Self-checking and rewriting against hedging and AI-speak (stripping "maybe", "perhaps", "it depends")

08Six-dimensional trace-back across references/research (books, dialogues, expressive DNA, third-party perspectives, decisions, timeline)

3External Dependencies

Kind	Dependency / Endpoint
cli	宿主 WebSearch / 联网检索工具（Step2 张雪峰式研究必需；由宿主提供，非 skill 自带脚本）

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 directives and rule text; contains no script execution or external impacts

Audit Notes: Grading basis (evaluating only this skill directory): examples/zhangxuefeng-perspective consists entirely of Markdown files (SKILL.md, FIDELITY.md, and 6 files in references/research/); contains no scripts, no binaries, no endpoints, no credential or environment variable access, and no disk write operations, classifying it as a "pure instruction" tier rated green. Non-directory aspects to note (do not change rating tier): Step 2 relies on host WebSearch/network tools (host capabilities, not code within this directory); ingestion of external content constitutes a prompt-level injection surface (see injection_surface); the pricing and venture investments of Fengxue Weilai described in SKILL.md represent documented background context, not runtime behavior. Free of danger flags, guardrail circumventions, and arbitrary code execution.

5Second-Pass Verification

[ok] 角色扮演规则与「已去世」前提属实 — SKILL.md 角色扮演规则段确有「直接以张雪峰的身份回应」「免责声明仅首次激活时说一次」「不跳出角色做meta分析」「张雪峰已于2026年3月24日去世，角色扮演基于其生前全部公开言论」，与 claim 一致；01-writings.md 也有同一事实的加粗句。
[ok] Agentic 三步与 Step2 四方向属实 — Step1 分类表（需要事实/纯框架/混合）、Step2 四个小节（看就业数据/看院校排名/看行业报告/看真实案例）、Step3 三条（先问家庭条件、引具体数据、给明确判断）均可逐条在内文找到，且含「必须使用工具（WebSearch等），不可跳过」。
[ok] CHECKPOINT 开口前三问与反例黑名单一致 — 「🔴 CHECKPOINT · 开口前三问」列出数据查了吗/第一句给判断了吗/家庭条件问了吗；反例黑名单 8 条亦重申「没问家庭条件就给追随热爱」「4 段铺垫后才给结论」「hedging 是 AI 腔不是东北哥」，两处口径吻合。
[ok] 5 模型 + 8 启发式与 FIDELITY 描述互证 — 正文核心心智模型确为 ①社会筛子论 ②选择>努力 ③就业倒推法 ④阶层现实主义 ⑤争议即传播；决策启发式确为 8 条。FIDELITY.md 写「心智模型5个（各含证据+应用+局限）、诚实边界6条、内在张力5对、反例黑名单8条、失败模式Fallback树9条」，计数一致。
[ok] §3 风险扫描 63 条 URL 的性质归类 — 抽样命中（jiemian、lifeweek、36kr、chinaventure、21jingji、lanjinger、thepaper、bilibili、guancha、wikipedia、sina/163/sohu 语录页等）全部位于研究文件的「来源索引」或正文链接引用，属文档化引用；无运行时 API 端点、无 webhook；凭证/执行/写盘/网络请求/危险旗标/沙箱六类均 0 命中。
[ok] 目录内无脚本 / 无凭证读取 / 无写盘 — §1 全部 8 个文件类型均为 text，§3「执行外部程序/写盘/网络请求/危险旗标/沙箱」五类均 0 命中；SKILL.md 仅以自然语言要求宿主使用 WebSearch，无本目录代码可执行。

6Verdict & Scenarios

High persona fidelity and complete expressive DNA: Northeast big brother tone, short sentences with high certainty, punchy joke-driven rhetoric, and comprehensive taboo blacklist, backed by an independent dual-agent evaluated FIDELITY.md (97/100, style fingerprint: "blind read 3 sentences to identify the person").

Data-first agentic architecture: Enforces WebSearch for employment rates, median salaries, and cutoff scores; explicitly states for out-of-scope questions that "I don't have the latest data, and giving advice from memory would be lying to you," serving as a benchmark for documenting persona skill information limitations.

Sharp, actionable realistic analytical framework: Social Sieve Theory, reverse employment deduction, Median Principle, Soul Inquiry, and family background triage apply directly to college application and career decisions without empty moralizing.

Anti-mechanization and boundary awareness: CHECKPOINT 3-question pre-speech gate, 8 counter-example blacklists, 9 failure mode fallback trees, and clear honesty boundaries (covering tensions between business ventures and educational philosophy, public persona vs. behind-the-scenes realities).

Best for: Ideal for students and job seekers from ordinary family backgrounds seeking employment-oriented Gaokao college application counseling and career planning: delivers actionable assessments on majors, colleges, and graduate school exams, requiring scores, province, and family financial circumstances before proposing solutions. Also serves as a design benchmark for "high persona fidelity + data-first" character skills—featuring a complete six-piece suite: persona rules, CHECKPOINT, counter-example blacklist, failure mode trees, expressive DNA, fidelity scorecard, and honesty boundaries.

Not for: Not suitable for educational research requiring rigorous academic debates or pluralistic viewpoints; not suitable for personalized planning for individuals from wealthy families pursuing academic, artistic, or non-profit paths (where employment-first advice might prove counterproductive); not suitable when authentic, citable original quotes directly from Zhang Xuefeng are required (outputs are simulations; citing them constitutes misattribution, and he passed away in 2026); also not suitable for environments lacking internet access for WebSearch.

Install Agent-Ready

① Verified Mirror 2026-10-10

Option A · Direct DownloadDownload zhangxuefeng-perspective.tar.gz

sha256: d57a9bbcfa95be6a…

Option B · Copy Guide for Agent

Agents automatically download and verify mirrors from this JSON guide.

② Upstream GitHub

GitHub Repo ↗

Pinned to commit fe03746870.