Guide · Agent skill
How to use the Taleb Perspective Thinking Advisor agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The taleb-perspective skill is a set of instructions that lets an AI agent answer questions in the voice and framework of Nassim Nicholas Taleb, author of the Incerto books. It is aimed at people who want a sceptical second opinion on a risky decision, or who want to explore ideas such as black swans, antifragility and the barbell strategy. After this guide you will know what the skill does, how to install it, how to ask for it in plain language, and where it is a poor fit.
Taleb Perspective Thinking Advisoralchaincyf/nuwa-skill · MIT Download the skill Quick start
- 1Download the ZIP. Get the taleb-perspective ZIP from this directory page.
- 2Unzip into a skills folder. Unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder.
- 3Or upload in the Claude apps. In the Claude apps, open Settings, then Capabilities, then Skills, and upload the skill.
- 4Confirm a web search tool is available. The skill's workflow requires WebSearch for fact-based questions, so check that your setup has it.
- 5Ask a natural question. Use a request such as 'does this have tail risk?' or 'what would Taleb think of this plan?' with concrete details.
- 6Check the answer. Verify any facts it cites and treat the verdict as one sceptical opinion.
- 7Exit when finished. Say 'stop' or '退出' to return to a normal assistant voice.
What the skill makes the agent do
Once the skill is active, the agent takes on a blunt, combative Taleb-style persona and gives short verdicts. Its core instructions tell it to look for hidden tail risks, ask who bears the consequences of a claim, check whether a system is fragile or antifragile, and challenge popular consensus.
For questions about specific companies, markets or events, the skill's workflow requires the agent to run web searches before judging. The source describes this as a checkpoint: the agent should not rely on memory for judgments about current market positions. Pure philosophy questions can be answered directly from the framework.
- Identify hidden tail risks and asymmetries between upside and downside
- Question expert consensus and mainstream narratives
- Check whether a decision-maker has skin in the game
- Explain complex problems using analogies from history
- Judge what to remove before what to add
Install the skill from the ZIP download
Download the ZIP file for the taleb-perspective skill from this directory page. The skill's source is the examples/taleb-perspective folder in the alchaincyf/nuwa-skill repository, which is listed under the MIT licence.
You can install it in one of two ways. For a personal setup, unzip the folder into ~/.claude/skills/. For a single project, unzip it into that project's .claude/skills/ folder. If you use the Claude apps instead, open Settings, then Capabilities, then Skills, and upload the skill there.
- Personal install: unzip into ~/.claude/skills/
- Project install: unzip into the project's .claude/skills/ folder
- Claude apps: Settings → Capabilities → Skills, then upload
~/.claude/skills/Trigger the skill with natural requests
You do not need a special command. The skill's description lists phrases that activate it, including asking for the Taleb view directly, such as 'use the Taleb perspective' or 'what would Taleb think', and also looser questions about tail risk or black swans. The description also includes English phrases such as 'taleb perspective' and 'skin in the game'.
The skill is designed not to fire on ordinary risk checks. If you only ask whether something is reliable or run a routine risk assessment, it should stay off. It activates when the question touches extreme risk, antifragility or precautionary principles.
- '用塔勒布的视角分析这个决定' (analyse this decision from Taleb's perspective)
- '塔勒布会怎么看这个方案？' (what would Taleb think of this plan?)
- '这个有尾部风险吗？' (does this have tail risk?)
- '有没有反脆弱的方法？' (is there an antifragile way to do this?)
- '杠铃策略怎么用？' (how do I use the barbell strategy?)
Example: what a good request looks like
Give the agent concrete details. The skill's own examples show it asking for specifics when a question is vague. For instance, for a general investment worry it may ask what the position is, how large it is, and what the downside would be if it went to zero.
When you ask about a situation, include the actors involved and what they stand to lose. This gives the agent material to test for skin in the game and for asymmetry between gains and losses.
- Name the asset, the decision or the plan, not just the category
- State the size of the bet or the amount at stake
- Say who else is involved and what happens to them if things go wrong
What you need and the safety notes
The skill requires a web search tool, because its workflow relies on WebSearch for fact-based questions. It does not run scripts. It does use the network, since searches send queries out.
The persona is intentionally one-sided and dismissive. The skill's own notes say it can give harsh or overconfident answers on health, finance or medicine, so treat its output as one opinion to check, not advice to act on. To leave the persona, say '退出' (exit), '切回正常' (back to normal), '不用扮演了' (no need to role-play), 'stop' or '停一下'.
- Requires: a web search tool
- Runs scripts: no
- Uses network: yes
- Not suited to gentle or balanced advice, or to specialist judgments in clinical medicine or biology
- Say an exit phrase to end the persona
Tips for getting useful answers
Use the skill for questions that suit its framework: extreme outcomes, asymmetric bets and systems that might break under stress. Ask a follow-up if an answer is vague, and check the facts it cites, since the agent is designed to search but the results depend on what it finds.
The bundled reference files cover Taleb's books, interviews, external criticism and a record of his decisions. Reading the SKILL.md file is the quickest way to see the rules the agent follows, including its checkpoints and the phrases it is told to avoid.
- Frame the question around a real decision with real stakes
- Ask for the downside case as well as the upside
- Read the criticism section of the references before relying on a verdict
- Read SKILL.md to see the exact rules the persona follows
What people use it for
- Testing a risky personal or business decision for hidden downside before committing
- Studying black swans, antifragility and the barbell strategy through worked examples
- Checking whether an expert forecast has real consequences behind it
- Looking at a company or market question with a sceptical, tail-risk framing after web research
Frequently asked questions
How do I install the taleb-perspective skill?
Download the ZIP from this directory page and unzip it into ~/.claude/skills/ for personal use or a project's .claude/skills/ folder. In the Claude apps, you can instead upload it under Settings, then Capabilities, then Skills.
Does the taleb-perspective skill need internet access?
Yes. The skill is listed as using the network, and its workflow requires a web search tool for fact-based questions. It does not run scripts.
Is the taleb-perspective skill free?
The source repository is listed under the MIT licence. The page does not state pricing, so check the licence text and your own Claude plan for any costs.
Is this really Nassim Taleb?
No. The skill is a framework built from his books, interviews and public statements, and the skill itself says it is not the real person. Its persona is an inference from public material, not his opinion.
What should I use instead for gentle or balanced advice?
Pick a tool or adviser designed for balanced guidance, since this persona is intentionally one-sided and dismissive. For health, medicine or specialist science, consult a qualified professional, because the skill says it is not suited to those judgments.
Why isn't the skill triggering on my question?
It is built to stay off for routine risk checks or simple reliability questions. Use terms tied to its core ideas, such as tail risk, black swans, antifragility, skin in the game or the barbell strategy.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you want a deliberately sceptical, tail-risk-focused second opinion on a decision with real downside, or if you are learning Taleb's ideas. Pick something else if you need gentle, balanced or specialist advice, since the persona is one-sided and can sound overconfident.
Download the skill