Put your expertise to work.
A rig turns what you know into an agent. Run it yourself, or share it with your whole team.
The old way
You ask Claude “build me a stock analysis tool.” You get a React dashboard with a dozen dependencies, a backend you'll never touch, and a deploy pipeline you don't understand.
It works for the demo. Then your process changes and you're debugging someone else's app. The tool became the problem.
The agent already exists
Claude can already read files, search the web, call APIs, and reason through problems. What it's missing is your process, your data, and your judgment.
A rigis a folder that gives an agent your brain. You don't write code. You write down what you know, and the agent does the rest.
Four folders, four agents that already know your work.
Your analysis framework, encoded
Your CLAUDE.md describes how you think: DCF weighted 60%, momentum 40%, always check insider filings first. Your data/ folder holds your watchlist and theses; your skills encode your morning routine.
Point Claude at the folder and it's your analyst, not a generic chatbot.
It starts as a folder.
Open it in any harness, your agents get to work, and your whole team shares one live context. That's a rig.