Use cases

🌱 Accessing restricted sites (eg Reddit) 🌱 Claude Code Mobile Remote Control

Other tips

45 Claude Code Tips: From Basics to Advanced Github Source You’re using Claude Code Wrong - These 10 Tips Will Change Everything YouTube

Agent Teams

This is a new multi-agent architecture where multiple Claude instances working in parallel, sharing a task list and with the ability to communicate to each other. The latter reduces the risk of sub-agents, that cannot not coordinate directly. However, this comes with higher token cost. Source

Uses of Claude Agent Teams Agent teams are overkill for small bounded work that does not require co-ordination, e.g tasks in which the agents simply feed a β€˜main’ agent. Use them instead for tasks where consistency is required across parallel deliverables or when building complex software in which different layers must co-ordinate. Source

Key Concepts of Claude Code

According to this YouTube video, the key concepts of Claude Code are?”

  1. What is Claude Code β€” an AI that takes actions, not just gives advice
  2. The Terminal β€” the black screen where Claude Code runs
  3. Prompts β€” plain English instructions you give Claude
  4. Permissions β€” controlling what Claude can do without asking
  5. Tool Use β€” built-in capabilities (read, write, bash)
  6. Context Window β€” Claude’s short-term memory
  7. Conversation History β€” saved sessions you can resume
  8. Token Usage β€” how costs are calculated
  9. Claude.md β€” your instruction manual file for Claude
  10. Memory β€” auto-saving preferences across sessions
  11. Compact Context β€” summarizing long conversations to free up space
  12. Models β€” Haiku, Sonnet, and Opus (speed vs. power tradeoffs)
  13. Denying Access to Files β€” blocking sensitive files via settings.json
  14. Flags β€” launch options that customize Claude’s behavior
  15. Extended Thinking β€” Claude’s built-in reasoning before acting
  16. Slash Commands β€” shortcuts like /compact, /clear, /help
  17. Skills β€” expert instruction playbooks for specific tasks
  18. Hooks β€” automated scripts that trigger at specific moments
  19. MCP Servers β€” connecting Claude to external tools like Notion or Airtable
  20. Sub Agents β€” specialists running in their own separate context windows
  21. Agent Teams β€” multiple agents that communicate directly with each other
  22. Multimodal Support β€” uploading images/screenshots to Claude
  23. Checkpoints β€” automatic snapshots before every file edit
  24. Git Integration β€” version control for tracking and rolling back changes
  25. CLI/Headless Mode β€” running Claude autonomously with no human input (-p flag)
  26. Cost/Pricing β€” Claude Max subscription vs. pay-as-you-go API
  27. Work Trees β€” multiple Claude instances working on isolated tasks simultaneously