Hey, it's Cindy 🌱 You commented files, so here is the whole structure from the reel as an implementation checklist, so Claude can find the right information without loading everything every time.
btw the waitlist for AI System Club is now open! If you're keen on using AI to automate your content, video editing, job search, business or build AI agents, feel free to sign up (only if you want to hehe). Ps. the waitlist gets first dibs:)
the outcome
What you get
A reviewable workspace map with a concise root CLAUDE.md, scoped project instructions, indexed memory, reusable skills, safe integrations, scripts, and version control.
Use the guide in order. Gather the real inputs, run the prompts against those inputs, inspect the evidence, then make one reviewable change at a time.
start here
Set it up step by step
Run the inventory prompt in read-only mode. It must report what exists, what loads, and what is missing before creating files.
Create the minimum root structure: CLAUDE.md, brain index, now.md, project folders, and one context file per active project.
Move long procedures into skills and deterministic repeated work into scripts. Keep the always-loaded root instructions concise.
Record MCP connections and tool ownership without putting credentials in documentation.
Initialize Git, add a safe .gitignore, scan tracked history for secrets, and verify the setup with /memory, /skills, /mcp, /hooks, and /doctor.
copy these
The prompts
Audit and build the 20-point structure
📋 copy this
Audit this workspace before editing. Report pass, missing, duplicate, or risky for: root CLAUDE.md; business, offer, customer, voice, and goal context; one workspace root; brain folder; brain index; memory folder; now.md; one folder per active project; scoped project instructions; model-routing rules; tools and connection inventory; central skills folder; workflows separated from skills; deterministic scripts for repeated tasks; Notion connection for open tasks; notes connection only if actually used; Git repository; secret files kept out of Git; and .env covered by .gitignore. Use /memory, /skills, /mcp, /hooks, /context, and /doctor to verify what Claude Code actually loaded. Propose the smallest file tree and exact diffs. Keep root CLAUDE.md under 200 lines where practical, move detailed procedures into skills, never copy credentials into docs, and do not write until I approve the plan.
check the work
Quality gate before you trust it
1 /memory shows the intended root and scoped instruction files.
2 /skills and /mcp show only the integrations needed for the current workspace.
3 The root CLAUDE.md is concise; detailed or occasional procedures live in skills.
4 now.md contains current priorities only and is not used as permanent memory.
5 Git status is clean of .env files, keys, tokens, generated secrets, and private exports.
6 Save the source links, as-of date, and final evidence beside the output so another person can reproduce the check.
Guardrail. There is no universal perfect folder tree. Use only the parts that match the real workflow, and never create empty bureaucracy for its own sake. Secrets belong in environment storage, not context files, memory, screenshots, or Git.
keep it useful
Save the evidence with the workflow
Save the prompt with the source links and the date you verified them. When the tool, policy, price, interface, or documentation changes, update the guide before you reuse the workflow. Keep the inputs and outputs beside the project so the next run can be checked instead of reconstructed from memory.
the waitlist is now open
AI System Club is launching soon
Build AI systems with me through recorded lessons you can follow at your own pace. Join the waitlist for first access and the founding price.
btw this list gets first dibs the moment doors open :)