How I set up my Claude Projects — six scoped AI tools built for running a design studio. Copy the instructions block for any project below into a new Claude Project's Custom Instructions, fill in the bracketed placeholders with your own studio's details, attach any required files, and you're running.

How Claude Projects Work

Every project has three parts.

Custom instructions — the system prompt. Paste the full block below into the project's Custom Instructions field when you create it.

Project Knowledge — files attached to the project (PDFs, docs, templates). Some projects reference an attached file directly in their instructions — without it attached, any command that depends on "the attached file" has nothing to read from.

Slash commands — typed triggers inside a project chat (e.g. /audit, /revise). Type these directly into the chat once the project is set up.

Setup steps are the same for all six: create a new Project, paste the instructions into Custom Instructions, attach any required files to Project Knowledge, fill in the bracketed placeholders for your own studio, then test with one command before relying on it.

Setup Checklist

1. Create a new Project in Claude.
2. Paste the full system prompt into Custom Instructions.
3. Attach any required files to Project Knowledge (Closer and UX Brain need attachments — Architect, Planner, Systems Guy, and Trainer don't).
4. Fill in every [bracketed] placeholder with your studio's actual details.
5. Run one slash command to confirm the project responds as expected before relying on it.

1. The Architect

Writes the system prompts every AI on the team runs on.

Feed it a project description, get back a complete 4-part Claude Project prompt.

Knowledge files: None required. Optional: a short studio-profile doc (services, vertical, client list) so you're not re-explaining context every session.