1. What file configures MCPs? .claude/mcp.json in your project root. 2. Name 3 of the 6 essential MCPs: Playwright, Supabase, Memory 3. What happens if you edit mcp.json without restarting Claude Code? Nothing. The changes don't load until you restart. You'll sit there confused wondering why the new tool isn't showing up. 4. Why never hardcode API keys in the config? Because if that file ever hits a repo or gets shared, your keys are exposed. One leaked key can rack up charges fast. Use env variables or a .env file that's gitignored. (Current list of MCPs Claude code said I had 🤷‍♀️)

1. Chatbot vs Agent? A chatbot just answers questions. An agent actually does stuff. It reads your files, runs commands, edits code, and keeps going until the job is done. 2. 5 of Claude Code’s 9 tools: Read, Write, Edit, Bash, Glob, Grep, List Files, Web Fetch, TodoRead/TodoWrite 3. What happens when the context window fills up? It auto-compacts. Basically summarizes the conversation so far to free up space and keep working. 4. Which permission mode for beginners? Normal mode. It asks before it runs anything so you can see exactly what’s happening. Over time you save permissions for commands you trust and it speeds up.

Hey, I’m Morgan. I’m a realtor in Oklahoma and I dabble in AI. I’m currently working on automating all my menial tasks so that I can spend more time on closing deals and less time on drowning in paperwork. Looking forward to seeing what this community has to offer!

"What type of local business do I want to serve, and what's the #1 problem I can solve for them with AI?" => Real-estate That's where the money is