原始内容
shuff.md
A personal Claude plugin that enables AI-assisted conversations with Patrick Shuff's perspectives on technology, startups, engineering, and life philosophy.
What is this?
This is a "digital twin" plugin - a structured knowledge base that allows Claude to answer questions as if you were interviewing Patrick Shuff. It's designed for:
- Interviewers who want to understand Patrick's thinking before or after a conversation
- Collaborators who want quick access to perspectives on common topics
- Anyone curious about technology philosophy, startup building, or engineering practices
Installation
Via skills.sh (Recommended)
Install directly from the skills marketplace using npx:
npx skills add patrickshuff/shuff.md
This will automatically detect your AI agent (Claude Code, Cursor, Codex, etc.) and install the skills to the appropriate directory.
As a Claude Code Plugin
Clone and install locally to your Claude plugins directory:
git clone https://github.com/patrickshuff/shuff.md.git ~/.claude/plugins/shuff
Then restart Claude Code or reload plugins. Skills will be available as:
/shuff.md:ask-shuff <question>/shuff.md:debate-shuff <position>
As a Standalone Project
Clone this repository and open it with Claude Code:
git clone https://github.com/patrickshuff/shuff.md.git
cd shuff.md
claude
Skills will be available without namespace:
/ask-shuff <question>/debate-shuff <position>
Usage
Quick Questions
Just ask naturally:
"What do you think about microservices?"
"How do you approach hiring?"
"What's your take on AI?"
Skills (Slash Commands)
When installed as a plugin, skills are namespaced:
/shuff.md:ask-shuff <question>- Get a response in Patrick's voice/shuff.md:debate-shuff <position>- Challenge a position and engage in debate
When used as a standalone project:
/ask-shuff <question>- Get a response in Patrick's voice/debate-shuff <position>- Challenge a position and engage in debate
Agents
For deeper exploration:
philosophy-explorer- Philosophical deep divesstartup-advisor- Startup-specific guidancetech-explainer- Technical explanations and opinions
Structure
shuff.md/
├── .claude-plugin/
│ └── plugin.json # Plugin manifest (required)
├── CLAUDE.md # Main context file
├── README.md # This file
├── AGENTS.md # Agent guidelines
├── commands/ # Legacy command definitions
│ ├── ask-shuff.md
│ └── debate-shuff.md
├── skills/ # Modern skill definitions
│ ├── ask-shuff/
│ │ └── SKILL.md
│ └── debate-shuff/
│ └── SKILL.md
├── agents/ # Agent definitions
│ ├── philosophy-explorer.md
│ ├── startup-advisor.md
│ └── tech-explainer.md
└── context/ # Knowledge base
├── philosophy.md # Core tech philosophy
├── startups.md # Startup perspectives
├── engineering.md # Engineering views
├── principles.md # Mental models
├── bio.md # Background info
└── topics/ # Deep-dive topics
├── unix-philosophy.md
├── interviewing.md
├── build-vs-buy.md
├── monorepo-vs-polyrepo.md
├── monolith-vs-microservices.md
├── internal-tooling.md
├── language-choice.md
└── ... (20+ more topics)
Customization
To create your own version:
- Fork this repository
- Update
context/bio.mdwith your background - Modify the philosophy files to reflect your views
- Fill in the topic placeholders with your perspectives
- Adjust the interaction style in
CLAUDE.md
Philosophy
This plugin embodies a few key ideas:
- Opinions matter - The value is in having clear perspectives, not neutral summaries
- Context enables quality - Rich background context enables better responses
- Structure aids retrieval - Organized knowledge is more accessible knowledge
- Placeholders are prompts - Empty sections with good prompts guide future content
- Everything is subject to change - Views evolve. That's the beauty of keeping this in source control. The git history is the record of how thinking changes over time. Commit often, update freely.
Contributing
This is a personal knowledge base, so contributions are mainly:
- Bug fixes to the structure
- Suggestions for additional topics
- Improvements to the skill/agent definitions
License
MIT