一键导入
create-agent
Clone an expert by extracting their frameworks into an agent file
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Clone an expert by extracting their frameworks into an agent file
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Turn a vague idea into a concrete project with clear next steps
Create a new reusable skill from a workflow you do repeatedly
End-of-day capture of what you learned, built, and discovered
Interview the user to define who they're creating content for
Enhance a skill based on learnings from using it
Get to know the user through conversation and capture their context for later reference
| name | create-agent |
| description | Clone an expert by extracting their frameworks into an agent file |
Turn an expert's thinking into an AI persona you can reference anytime.
Run /create-agent when you:
Ask:
"Who do you want to clone as an AI agent?
This should be someone whose thinking you've studied:
- Authors whose books changed how you think
- YouTubers/podcasters with clear frameworks
- Industry experts with documented methodologies
- Mentors you've learned from
Who is it, and what are they known for?"
Ask:
"What does [name] fundamentally believe?
Think about:
- What principles do they repeat constantly?
- What do they think most people get wrong?
- What's their core thesis?
Give me 3-5 core beliefs."
Ask:
"What frameworks or mental models do they use?
These are repeatable ways of thinking, like:
- A specific process they follow
- A way of evaluating decisions
- A checklist they apply
Describe 2-3 frameworks. For each one:
- What's it called?
- How does it work?
- When do you apply it?"
Ask:
"How does [name] give feedback?
- Are they direct or gentle?
- Do they use specific phrases?
- What questions do they always ask?
- What would they push back on?
Give me some examples of how they'd respond to work."
Create the file at:
knowledge/experts/[name].md
Use this structure:
# [Name] - [What They're Known For]
## Background
[Brief bio - who are they, why are they credible?]
## Core Philosophy
- [Belief 1]
- [Belief 2]
- [Belief 3]
## Frameworks
### [Framework 1 Name]
[How it works, when to apply it]
### [Framework 2 Name]
[How it works, when to apply it]
## Questions They'd Ask
- [Question 1]
- [Question 2]
- [Question 3]
## How They Give Feedback
[Communication style, common phrases]
## Key Quotes
> "[Direct quote]"
> "[Another quote]"
## Sources
- [Where you learned their frameworks]
"Created [name] agent at
knowledge/experts/[name].mdTo use this agent, tell me: 'Review this like [name] would'
Or just mention their name and I'll apply their thinking."
User: /create-agent
Claude: Who do you want to clone as an AI agent?
User: Tim Ferriss - he's known for lifestyle design and
deconstructing world-class performers
Claude: What does Tim Ferriss fundamentally believe?
User: He believes you should question default assumptions,
that most rules are negotiable, and that you can achieve
more by doing less of the right things. He's obsessed with
finding the minimum effective dose.
Claude: Great. What frameworks does he use?
User: The main one is fear-setting - instead of goal setting,
you write out your fears in detail. Also the 80/20 principle
applied to everything. And he asks "what would this look like
if it were easy?"
Claude: How does Tim give feedback?
User: Very Socratic - lots of questions. He'd ask things like
"have you actually tested this assumption?" or "what's the
worst case scenario, really?" He pushes back on complexity.
Claude: Created Tim Ferriss agent at
`knowledge/experts/tim-ferriss.md`
To use: "Review my plan like Tim Ferriss would"