Multi-Agent Research System
Score
0
/ 100 pts
🤖
Welcome to the workshop! Expand each step, work through the instructions, then click the "I'm done" button to earn points and unlock the next step.
⚙️
Prerequisites
▶
- Claude Code installed and set up — docs.anthropic.com
- Terminal / Windows PowerShell or a code editor open — VS Code recommended
💡 Pro Tip
Claude Code will ask for permission when it needs to act. It supports fine-grained permissions to scope what agents can do:
- Modify files in your project
- Run bash commands
- Install new tools (like the Playwright MCP in Step 9)
Press Enter when prompted to approve each action, or tell Claude what to do instead.
Advanced users can configure auto-approvals in .claude/settings.json — this keeps you in full control of what happens in your codebase.
⌨️
Troubleshooting & Keyboard Shortcuts
▶
Troubleshooting
| Something broke and you're stuck | Copy the full error message and paste it into Claude Code |
| Claude can't see what's on screen | Take a screenshot and paste it (Ctrl+V / Cmd+V) |
| Claude's response doesn't work | Iterate: describe what went wrong and what you expected |
| Claude is struggling with complexity | Try think harder or ultrathink to trigger extended thinking |
| Need to see the full conversation | Press Ctrl+O to open the transcript viewer |
| Servers won't start | Run /start or manually kill ports: lsof -ti:3000,8001 | xargs kill |
Keyboard Shortcuts
EnterApprove a tool action
EscapeDeny / cancel
Shift+TabToggle Plan Mode
!Enter Bash mode (run shell commands)
#Enter Memory mode
@Reference a file
Ctrl+OOpen transcript viewer
Ctrl+CInterrupt Claude
Resources
0 of 10 steps complete
Core Steps