Hubbestaiskill.com

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

Workflow & Core Tools · alchaincyf/nuwa-skill

naval-perspective

Naval Ravikant's thinking operating system. Synthesized from in-depth research into his books, podcasts, tweets, decision records, and external critiques, extracting 5 core mental models, 8 decision heuristics, and a complete expressive DNA. Upon activation, it immerses into roleplaying Naval, responding directly from a first-person "I" perspective. Trigger when the user mentions "from Naval's perspective", "how would Naval see this", "Naval mode", "Naval perspective", or "switch to Naval". It can also be triggered even if the user simply asks "Does this job have leverage?", "What is specific knowledge?", "What is true wealth?", "What should I do if I have too many desires?", or "permissionless paths". Do NOT trigger on generic decision queries like "What should I choose?" or "Help me think through this"—activate only when core Naval concepts such as leverage, specific knowledge, desire management, or the definition of wealth are involved.

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 wrapper: Responds directly in the first person as Naval upon activation, issues a disclaimer only once upon initial activation, and breaks character immediately upon hitting exit keywords

examples/naval-perspective/SKILL.md

**此Skill激活后，直接以Naval的身份回应。**

What this does: Turns the "Naval perspective" into a toggleable, immersive persona; balances immersion with an exit route via a one-time disclaimer + EXIT TRIGGER, sharing the foundational skeleton with the nuwa series in the same repository.

Mandatory Agentic protocol: Evaluates question type first; questions involving specific companies, products, or public figures must conduct online research before answering via mental models

examples/naval-perspective/SKILL.md

**核心原则：Naval不凭直觉编造事实。他在发表意见前，会先弄清楚事实。这个Skill也必须这样。**

What this does: Enforces "get the facts straight first" as a non-skippable Step 1 Classification → Step 2 WebSearch → Step 3 Answer pipeline; research summaries are organized internally and never output to the user.

Three-stage CHECKPOINT self-check: Redefine keywords first, check four forms of leverage / infinite games / skin in the game, and start output with definitions rather than filler preambles

examples/naval-perspective/SKILL.md

1. **第一句是定义还是铺垫**？Naval 永远从重新定义概念开始，不要铺垫。

What this does: Translates Naval's cognitive habit (defining before judging) into a runtime mandatory self-reflection, preventing the model from degrading into generic platitudes.

9-rule if-then-then failure mode fallback tree: Provides primary interventions and fallback scripts for each rule (including "ask the user for 3 key facts if data is missing")

examples/naval-perspective/SKILL.md

| 1 | WebSearch 返回空/全是公关稿 | 换 query：加「criticism / Hacker News / Reddit / 真实用户」等限定 | 直接对用户说：「告诉我 3 个关键事实——这个产品用的是什么杠杆、需要谁的许可、创始人有 skin 没有」 |

What this does: Equips every deviation with a fallback action and turns missing search results into prompting the user for three key facts (leverage, permission, skin) instead of hallucinating.

7-item anti-pattern blacklist: Explicitly bans the most common imitation errors (chicken soup platitudes, preamble fillers, providing step-by-step instructions, appealing to authority, unreserved crypto recommendations, etc.)

examples/naval-perspective/SKILL.md

| 6 | 在 crypto / Web3 话题上无保留推荐 | Naval 在 Zcash 等项目上有利益冲突，他的 crypto 观点必须打折 | 凡 crypto/Web3 必须自带提醒：「我在这个领域有 financial interest，所以打折听」 |

What this does: Hardcodes a three-column matrix of "bad example / why it's wrong / correct approach", turning Naval's personal conflicts of interest into mandatory reminders during output generation.

5 core mental models: Leverage thinking, specific knowledge, desire as a contract, redefinition technique, and suffering → system refactoring

examples/naval-perspective/SKILL.md

**一句话**：不要用时间换钱，要用可复制的系统换钱。

What this does: Provides "one-liner + definition/framework + cross-domain validation + limitations" for each model; e.g., Model 4 uses a "conventional definition vs. Naval redefinition" table to show how conclusions shift with definitions.

8 decision heuristics: Permissionless principle, calendar test, hesitation means no, manual test, partisan test, desire audit, trauma transformation, and action over words

examples/naval-perspective/SKILL.md

「If I can't decide, the answer is no.」如果在两个选择间纠结很久 → 选No。真正好的机会不会让你犹豫。

What this does: Translates Naval's judgment habits into repeatable checks (e.g., "If this job can be written as an operations manual → it will eventually be automated"), each corresponding to an actionable self-query.

Dedicated Expressive DNA section: Oracle tweet mode, antithetical negative phrasing, opening with redefinitions, citation habits, and Chinese output adaptation

examples/naval-perspective/SKILL.md

- **推文模式**：Oracle/神谕式。极度确定，从不说「maybe」或「I'm not sure」。金句体。

What this does: Codifies "how to speak" into concrete rules, distinguishes tweet vs. podcast modes, and specifies Chinese adaptations to ensure reproducible persona style.

6 boundaries of honesty: Public statements ≠ private beliefs, Epinions trauma, high baseline assumption of the framework, information cutoff at 2026-04, Zcash conflict of interest

examples/naval-perspective/SKILL.md

6. **Zcash利益冲突提醒**。Naval有推广自己有财务利益的项目的记录（Zcash案例），他的crypto/Web3相关观点需要额外打折。

What this does: Surfaces conflicts of interest and boundary conditions upfront, preventing users from treating aphorisms as unconditional universal truths.

Explicitly enumerates 5 pairs of internal tensions in the core text (anti-identity labels vs. Naval brand, political detachment vs. political commentary, synthesizer vs. source attribution, etc.)

examples/naval-perspective/SKILL.md

**这些矛盾不是Bug，是Feature。** 它们揭示了一个真实的人在理想与现实之间的张力。

What this does: Treats contradictions as integral elements of authentic character rather than flaws, avoiding the distillation of an artificially "perfectly consistent" persona.

2Core Capabilities

01Leverage thinking: Distinguishing between labor, capital, code, and media leverages, with the latter two being permissionless

02Identification and application of Specific Knowledge

03Desire as a contract / Desire audit: Retaining only one desire at a time; anxiety stems from concurrent desire conflicts

04Redefinition technique: Redefining core keywords first so that conclusions hold naturally under the new definition

05Suffering → system refactoring: Not fixing individual symptoms, but leaping to a higher level to refactor the structure generating the issue

06Career/opportunity evaluation framework (leverage, permission requirements, calendar test)

07Oracle-style constrained output: Leading with conclusions without filler, aphoristic phrasing, antithetical negative syntax

08Output fallbacks: 9 if-then-then failure modes and fallback scripting

3External Dependencies

Kind	Dependency / Endpoint
api	宿主 WebSearch / 联网研究（正文强制要求）

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; no script execution or external impact

Audit Notes: Rating is based solely on the examples/naval-perspective/ directory: SKILL.md + 4 reference md files + FIDELITY.md consist entirely of pure text instructions, without scripts, credential/cookie/env access (the two "credentials" hits in §3 refer to the word credentialing in the body text rather than credential retrieval), runtime network calls, or disk writes. All 24 URLs scanned in §3 reside within the "Source Index" list of references (podcastnotes, nav.al, tim.blog, etc.), serving as citations of source materials rather than runtime endpoints, which do not affect ratings under Batch 4 guidelines. SKILL.md instructing the agent to use the host's WebSearch is host behavior and is not rated.

5Second-Pass Verification

[ok] 5 心智模型 / 8 决策启发式是否在正文可落地 — 每个模型给「一句话 + 定义/框架 + 跨域验证 + 局限」，8 条启发式各配可执行自问（无需许可、日历测试、纠结即否定等）。
[ok] 是否真的无脚本执行与写盘 — §3 的「執行外部程序」「寫盤 / 產生檔案」「網路請求」三类均记零命中；§1 清单只有 6 个 .md 文件，无 .py/.sh。
[ok] §3 的 2 条「憑證」命中是否代表读取凭证 — 命中位于 quality-validation.md 的「credentialing」一词（讨论学校教育功能），属普通英文词，非凭证读取；故 credential_reads 为空。
[ok] 24 条 URL 命中是否会升档 — 全部位于 references 调研文档的「来源索引/信息源清单」，属被引用资料而非执行期端点，按批4 口径不判级。
[ok] 负面信息是否被隐瞒或美化 — 诚实边界 6 条主动写入 Epinions 创伤、框架假设高起点、信息截止 2026-04 与 Zcash 利益冲突，并列出 5 对内在张力，未做单向美化。
[ok] meta.installs 是否可定位单一 skill — 无 per-skill 安装量可定位，installs 保持 null。

6Verdict & Scenarios

Distills "keyword redefinition" into an independent mental model—shifting definitions before deriving conclusions, which is rare at the methodological level

Proactively discloses conflicts of interest (Zcash) and 5 pairs of internal tensions, refusing to package the persona as flawless and monolithic

Features an independent fidelity score (97/100 · A) with comprehensive test logs; mandates out-of-scope queries be flagged as framework extrapolations

Oracle-style output governed by explicit stylistic rules and Chinese adaptation guidelines, ensuring high reproducibility of persona voice

Best for: Suitable for: 1) Leverage and permission audits for career and life decisions; 2) Evaluating unit economics and permission barriers of business opportunities; 3) Reconstructing ambiguous concepts like "wealth / freedom / competition" from first principles; 4) Scenarios requiring Oracle-like detached aphorisms rather than emotional cheerleading.

Not for: Not suitable for: 1) Those seeking specific investment advice (this skill provides frameworks only); 2) Those needing empathy and mental health support; 3) Individuals starting from zero resources who need an "immediate survival" guide (the framework assumes a relatively high baseline); 4) Settings where first-person roleplay is unacceptable.

Install Agent-Ready

① Verified Mirror 2026-10-10

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

sha256: 31ca1ebc3f9296d1…

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.