Claude.ai vs Claude Code: Stop Chatting and Start Commanding (Hands-On Workshop)

Most of us still chat with Claude. This hands-on workshop shows you how to command it instead. In this session, Mansi Mutreja (AI & ML Engineer, Enqurious) walks through what it actually takes to go from casual prompting to using Claude Code as an agent that reads your files, fixes your bugs, runs your tests, and ships work for you. We build it all on a real Python project (Orbit Watch, a satellite-telemetry system), so every concept is shown live, not on slides. What you'll learn: The real difference between Claude.ai (a chatbot) and Claude Code (an agent that takes action) The four pillars of a Claude Code power user: context, skills, execution mechanics, and model choice How to write a CLAUDE.md that actually helps (keep it short: project overview, tech stack, conventions, and the things Claude should never do) Permissions as supervised delegation: low, medium, and high-risk actions, and how one line turns an autonomous agent into a supervised one CLAUDE.md vs memory.md, and when each one applies Built-in skills (resume, clear, compact, doctor, cost, usage) and how to build your own custom skills Running a multi-agent task: a supervisor delegating to sub-agents end to end Practical token-saving habits: reference files by path, edit instead of following up, pick the right model (Haiku, Sonnet, Opus), and skip the "please" and "thank you" Whether you are a developer, QA engineer, PM, or founder, you'll leave knowing how to make Claude do real work for you, safely. A recommended CLAUDE.md article and a skills list are shared after the session. #ClaudeCode #Anthropic #AIagents #PromptEngineering #Enqurious