Claude Code & Cowork – The Full Masterclass

Explore Claude Code permission modes: accept edits, ask permission, auto mode, and bypass. Learn how these settings govern when Claude edits, tests, and runs tools to build a browser game.

Select a model based on features and credits, with Sonnet as the default for most tasks, and adjust effort and the context window to balance quality and usage.

Master quick Claude Code workflows by cueing messages, setting a project folder and bypassing permissions, managing sequential and parallel tasks, and using ctrl z, ctrl y, dictation, and side chat.

Learn to craft clear, well-structured prompts for Claude by outlining goal, context, changes, what not to change, and verification to achieve precise results.

Plan mode lets you plan without editing files, exploring directions for a complex feature before implementation. Add comments, revise the plan, and approve it before Claude executes in auto mode.

Use screenshots and image files to rapidly communicate UI changes, swap icons, and derive design inspiration for buttons and layouts, using SVGs and reference images to guide style in prompts.

Use the goal feature to manage comprehensive, end-state tasks across longer sessions, employing a faster model to verify progress against a clear end condition and ensure long-term consistency.

Explore slash commands in the Claude Code masterclass, including usage and model changes, and learn to manage context, tokens, and custom commands via md files.

Explore how Claude memory files work, including user and auto-generated memory, memory.md indexing, and project asset conventions; learn to consolidate memory and load memory.md into the context.

Explore the compact command that frees context by summarizing conversations when the context window is full, saving tokens and guiding when to clear or reset.

Manage sessions in Claude Code with resume, fork, and rewind across cloud code, the desktop app, and the Visual Studio Code extension, and learn git checkpoints.

Paste the full error message into Cloud Code to scan key details, then use debug mode for rigorous diagnostics; if issues persist, switch to smarter models.

Set up a new town defense game project called the last village, plan its core architecture in plan mode, and showcase advanced features like sub agents and GitHub version control.

Audit and review your code using planned, visual, security, and general reviews. Learn to draft thorough prompts, prioritize fixes, and implement iterative improvements for a cleaner, safer codebase.

Refactor code safely in ongoing projects by removing redundancy, clarifying names, and improving structure, while verifying unchanged functionality through small steps and leveraging version control with Git and GitHub.

Learn how Git and GitHub track project history with commits, differentiate local and online repositories, and configure Git identity and GitHub CLI authentication to enable collaboration.

Set up a local git project by initializing a repository with git init, adding a readme file, staging changes with git add, and committing to establish a local history.

Create a GitHub repository, stage changes with git add, commit with git commit, and push with git push to publish your local project online.

Pull changes from GitHub to your local project using git pull, compare push and pull workflows, and see how Claude can automate commits and updates.

protect secrets by keeping api keys out of your code and away from public repos. use a .env file for sensitive values and a .gitignore to prevent commits.

Create and implement vector graphics as pngs or svgs into the game using assets generated with ai tools, cloud code, and automated cropping and naming; adjust gameplay directions as needed.

Learn how to run and preview your app quickly by using the preview button to start the preview server, load localhost, and refresh for immediate results.

Deploy your tower defense game from github to a live website with vercel. Learn the end-to-end workflow: cloud code builds, pushes to github, and vercel deploys updates automatically.

Learn how MCP (model context protocol) acts as a universal adapter and communication method—enabling Claude to connect to external tools such as databases, Google Drive, and GitHub.

Navigate the customize menu and the MCP slash command to connect apps like Gmail, Drive, and Calendar. Authenticate permissions, draft emails, and organize your inbox with labels and automated tasks.

Explore Claude skills, reusable workflows that automate prompts via markdown-based skills with metadata, enabling custom workflows, installation, and use for emails, newsletters, and presentations.

Skills are markdown md files; learn a better method to create, download, install, and restart Claude to use them across chat, co-work, and code.

Create custom skills directly in the interface or by turning a solved workflow, like a code review, into a reusable project or user-level Claude skill with description and trigger.

Learn how cloud code plugins package reusable skills and MCP servers, install and create them, and use the Anthropic marketplace to connect tools like Canva.

Understand subagents and how they work by building a main agent with parallel subagents that have separate context windows, enabling cost savings, specialized tools, and models.

Create sub-agents to run design and security reviews in parallel background tasks, using Opus for depth, with transcripts, and persist as custom sub-agents defined by YAML front matter.

Discover how hooks enforce non-negotiable rules in Claude code, from pre-tool to post-tool actions, gatekeepers, and how project, local, and global settings travel with the workflow.

Create and implement project hooks to prevent edits or deletion of files by wiring two JavaScript scripts and configuring settings.local.json, dot env, and gitignore in a Claude Code course project.