Build a Personal Claude Harness in Under an Hour: One Folder, One CLAUDE.md

@qendresahhotiCapturedClaude Code / Skills

Exact instruction

Create a root folder (e.g. 'my-os').
Add a me.md: plain text description of who you are, what you're working on, how you work.
Pick one domain to start (content, clients, studies) — create a folder for it and drop everything you have into it.
Write a CLAUDE.md that maps what lives where — a few sentences per folder is enough.
Open Claude Code in that directory and start asking questions. Claude will read the map, load the right context, and answer from your actual files.

Suggested prompt

Help me create a personal Claude harness: (1) set up a root folder with a me.md describing who I am and how I work, (2) create one domain folder with my existing files, (3) write a CLAUDE.md map of what lives where. Start with the me.md.

Adopt?

Yes: This is exactly the architecture already used in this project (CLAUDE.md + knowledge.md). The me.md addition is worth adopting for any personal knowledge project. General Claude Code update.

Find the resource

Creator @qendresahhoti has a GitHub starter harness — comment "SETUP" to get the link. Search: "qendresahhoti Claude harness github starter" to find it directly.

show original caption

How to build your own harness in under an hour. This is the system that makes Claude feel like it knows your whole business: a folder, one domain inside it, and a CLAUDE.md that maps what lives where. One folder. One domain. One map. That's the entire setup behind agents that seem to know everything. I put my exact starter on github, a working harness you can open and ask questions of before you change a thing. Comment SETUP and i'll send it to you. #ai #tech #developer #tools #claudeai