The exact setup that turns Claude from a forgetful contractor into a team member who knows your business, follows your rules, and runs jobs while you sleep. Scrubbed clean, packaged, and ready to install in about 30 minutes.
Every correction, decision, and preference becomes a memory the model recalls when it is relevant. Tell it once that someone left the company, or that you never use em dashes, and it stays told.
The longer you use it, the more it works like a teammate who has been with you for a year, not a contractor who forgets everything overnight.
Each layer is simple. They interlock so a single request draws on all of them at once.
No re-briefing tax. Every session starts informed by your constitution and memory.
Feedback becomes a memory file or a hook. It aligns to you the longer you use it.
The firewall blocks rm -rf /, force-push to main, DROP TABLE, reading .env - deterministically.
Skills carry their own reporting line, decision rights, and templates. Delegate a team member.
Scheduled jobs pull from your tools and deliver finished summaries and fixes overnight.
Run one command, answer a few questions, and it writes your personal config for you.
Run one command, then answer a few questions. Claude writes your CLAUDE.md, goals, memory seeds, and first skill, and shows you each file before saving. Nothing personal is written until you confirm.
$ bash bootstrap.sh ✓ Backed up existing ~/.claude ✓ Installed 4 hooks (firewall, format, ...) ✓ Installed /personalize-setup ✓ Safety core active ...mechanical install done. $ claude > /personalize-setup What's your name? Your orgs? Your stack? Top goals? -> writes your CLAUDE.md, memory, first skill.
Grab the kit, unzip it, and open START-HERE.md.
Run bash bootstrap.sh. It backs up your config and installs the safety core.
Open Claude Code and run /personalize-setup.
Your config, memory, and first skill are written for you.
Free for FFL members. Download it, unzip, open START-HERE.md, and let it interview you.