YouTube Deep SummaryYouTube Deep Summary

Star Extract content that makes a tangible impact on your life

Video thumbnail

Claude Code - 47 PRO TIPS in 9 minutes

Greg Baugues • 2025-05-22 • 9:19 minutes • YouTube

📚 Chapter Summaries (12)

🤖 AI-Generated Summary:

Mastering Claude Code: Pro Tips from a Developer’s Experience

Hey developers! If you’re diving into Claude Code, you’re in for a powerful coding companion. Greg, a fellow developer, recently shared his journey and some invaluable pro tips on using Claude Code effectively. Drawing heavily from Boris Churnney—the creator of Claude Code at Anthropic—Greg unpacks features and strategies that can elevate your coding workflow. Here’s a comprehensive guide based on his insights.

What is Claude Code?

Claude Code is an AI-powered coding assistant that operates primarily through a command-line interface (CLI). It’s designed to streamline your development process, automate tasks, and integrate seamlessly with your existing tools and workflows.

CLI Basics & Command Line Arguments

If you’re comfortable with bash or other CLI tools, you’ll feel right at home with Claude Code. Here are some essentials:

  • Passing Arguments on Startup: You can supply command line arguments that execute when Claude Code starts.
  • Headless Mode: Use -p to run Claude Code without the interactive UI.
  • Chaining and Piping: Claude Code can be combined with other command line tools and accept piped input.
  • Multiple Instances & Sub-agents: Run several Claude Code instances simultaneously. When you see “task” or ask Claude to spin up sub-agents, it’s actually launching additional instances.

This flexibility makes Claude Code powerful in automation and multitasking scenarios.

Working with Images & Screenshots

Images aren’t just static assets; Claude Code can interact with them dynamically:

  • Dragging Images: On macOS, drag images directly into the terminal to use them in Claude.
  • Screenshots: Use Shift + Command + Control + 4 to copy a screenshot, then paste it into Claude Code with Control + V (note: not Command + V).
  • Use Cases:
  • Mockups: Paste UI mockups and have Claude build interfaces based on them.
  • Iterative Feedback: Take screenshots of your app, feed them back to Claude, and get iterative improvements.

For automation, integrate Puppeteer (a headless browser tool) to capture screenshots programmatically.

Automation with Puppeteer & MCP Servers

Claude Code can automate browser actions via Puppeteer, capturing screenshots or simulating user interactions. Beyond that, Claude Code serves as both an MCP (Multi-Client Protocol) server and client, enabling communication with other agents and services.

  • Popular MCP Servers:
  • Postgres server for direct database integration.
  • API wrappers providing up-to-date documentation.
  • Dev tool companies like Cloudflare use MCP servers to supply live docs.

This connectivity allows Claude Code to fetch live data and interact with your development environment fluidly.

Fetching URLs & Working with Documentation

Claude Code can fetch and read URLs directly, making it perfect for:

  • Pulling in documentation from websites.
  • Building app logic based on external knowledge sources.

For example, Greg built a game for his daughter by feeding Claude the rules from unorules.com, bypassing the need to manually input or rely on pretrained data.

The Power of claude.md

One of the most crucial features is the claude.md file:

  • This markdown file is loaded with every request to Claude Code.
  • It contains essential instructions like coding style, bash commands, test running procedures, and repository etiquette.
  • Initialize it with /init in your project directory; Claude will scan and summarize the project structure for you.
  • You can create global or directory-specific claude.md files to tailor Claude’s behavior.
  • Keep your claude.md concise, specific, and regularly refactor to avoid bloat.
  • Use Anthropic’s prompt optimizer tool for refining your claude.md.

This file acts as your project’s "brain," guiding Claude to produce context-aware and consistent output.

Slash Commands: Custom Prompt Templates

Slash commands are customizable prompt templates you define in the cloud/commands folder:

  • Examples include commands for refactoring, linting, or reviewing pull requests.
  • They accept command line arguments, which are interpolated into the prompts dynamically.
  • Slash commands speed up repetitive tasks and enforce coding standards with ease.

UI Tips for Better Sessions

To maximize your interaction with Claude Code:

  • Tab Completion: Use tab to auto-complete filenames and directories, improving prompt specificity.
  • Interrupt Early: Don’t hesitate to hit Escape to stop Claude if it goes off track.
  • Undo Work: Use the undo feature to revert Claude’s previous action and steer it back on course.

Interrupting and correcting Claude early helps maintain productive sessions and prevents bigger errors.

Version Control Integration

Version control is your safety net when working with AI-generated code:

  • Frequently commit changes and ask Claude to write commit messages—it often crafts excellent, descriptive messages.
  • Revert often to previous stable points instead of trying to fix complex breaking changes.
  • Install GitHub CLI for smooth interactions, or use the GitHub MCP server as an alternative.
  • Claude can file pull requests and perform code reviews, helping automate parts of your development pipeline.

Managing Context & Memory

Claude Code has token limits and context windows to consider:

  • Monitor the auto-compacting indicator to know when context will be trimmed.
  • Compact context at natural breakpoints like after commits or task completion.
  • Sometimes clearing context and starting fresh leads to better results.
  • Use scratchpads or GitHub issues to help Claude plan without bloating conversation history.
  • For token-cost sensitive projects, externalize memory and carefully manage context size.

Managing Costs & Plans

Claude Code’s token usage can get expensive:

  • For teams, consider integrating open telemetry tools like DataDog for cost tracking and dashboards.
  • Upgrading to Claude Max plans ($100 or $200 monthly) bundles tokens and reduces per-token costs.
  • Greg noted spending about $150 in tokens over three days but found the cost worthwhile given the productivity gains.

Final Thoughts and Resources

Claude Code is a robust and flexible AI coding assistant with many hidden features and best practices. Greg encourages exploring Boris Churnney’s detailed post for even more pro tips and checking related resources for deeper understanding.

Ready to Level Up Your Coding with Claude Code?

If you want to supercharge your development with AI, start experimenting with these pro tips today. Use claude.md to set clear instructions, leverage MCP servers for live data, automate screenshots with Puppeteer, and keep your sessions tidy with thoughtful context management.

For the full post by Boris Churnney and other helpful links, check the video description or visit the official Claude Code documentation.

Happy coding! 🚀

Author: Greg, Developer and Claude Code Enthusiast
Special thanks to Boris Churnney and the Anthropic team for their pioneering work on Claude Code.

📝 Transcript Chapters (12 chapters):

📝 Transcript (229 entries):

## [00:00] Hey, my name is Greg. I'm a developer and over the last few months, Claude Code has become my default way of writing code. And so in this video, I want to walk you through some Claude Code pro tips. These pro tips are primarily based on this post written by Boris Churnney, who's the creator of Claude Code at Enthropic. And we're going to go through these uh pro tips pretty quickly. First tip, cla code is a ## CLI Basics & Command Line Arguments [00:24] CLI. So all the things that you're used to doing with other bashbased CLIs, you can probably do with cloud code. For instance, you can pass in command line arguments which will be run on startup. You can use -p to run it in headless mode. You can chain it with other command line tools. You can pipe data into it. You can run multiple instances of it at once. You can actually have cloud code launch instances of claude code. In fact, anytime you ask it to spin up a sub agent or anytime you see task, that's exactly what clog code is doing. Next category, images. You can ## Images & Screenshots [01:02] use an image simply by dragging it in to the terminal on OSX. You can use shift command control 4 to copy the screenshot and then use controlV to paste it into claude. That's controlV, not commandV like you're used to. There's two ways that you might find yourself using images a lot. The first is mockups. You can design a mockup, paste the mockup into Claude, and then ask it to build that interface. Second, you can use images to close the feedback loop with Claude. Ask it to build something, open up what it built, and then take a screenshot, feed that back into Claude, and it's pretty good at iterating when you're giving it feedback. Now, that's a manual process for taking images. You can also automate the screenshotting by using the Puppeteer MCP server, which is ## Automation with Puppeteer [01:54] pretty easy to set up and run locally. Then you can ask Claude to use Puppeteer to go open up the app, take a screenshot of it, and it can save those screenshots to your local directory. Speaking of MCP service, Cloud Code can function as both ## MCP [02:09] an MCP server and an MCP client. So that means that you can actually turn clogged code into an MCP server that can then be used by other agents. There's a whole bunch of MCP servers that you could use. It would be a whole video on its own just to go through some of the most popular ones. So we'll just hit a couple. For instance, you might find it useful to use the Postgres server to hook up Cloud Code directly to your database. You can use MCP servers that are effectively wrappers around APIs. Other dev tool companies like Cloudflare are using their MCP servers to provide up-to-date documentation to Claude. Not all dev tool companies are making their docs available via MCP just yet. So if you just paste in a link, Claude code ## Fetching URLs & Documentation [02:55] can fetch that URL and then use those docs to build against. You might also want to use fetch URLs to retrieve knowledge from the world that you use in your app. For instance, I built a game for my four-year-old daughter that was uh Bluey Uno. Instead of trying to describe the rules myself or relying on the training data for Uno rules, I pasted in unorules.com and had Claude code the gaming logic based on what it read ## Claude.md [03:23] there. Next category, claude.mmd. This is actually the first pro tip that's mentioned in Boris's post. A claw.md is a prompt that is loaded with every request that you make to claude code. This might include instructions for your project such as common bash commands to use, style guidelines, linting guidelines, how to run your tests, repository etiquette. If you type /init after you launch claude in a directory, it will create this claude. MD file for you after scanning the directory and summarizing its structure. If as you're coding you want to add instructions to the cloud.MD, you can use the hash sign. You can also set a global cloud. MD in your home directory/cloud. This will be loaded anytime that you're using cloud code across any project. You can also add a cloud.md file in subdirectories. You should also refactor your cla.md files often. So, it's common for them to grow in complexity as you continue to work on a project. But remember that this is a prompt that is being loaded on every turn of conversation with clawed code. And these models do much better the more specific you are. So, you don't want this to be crammed with a bunch of duplicative extraneous information. You can use Anthropic's prompt optimizer tool to help you write better cla.md files. Slash commands. You can define ## Slash Commands [04:56] these in thecloud/comands folder, and they're just prompts. So, for instance, here's one mentioned in Boris's post about solving GitHub issues. You might write a slash command for refactoring. You might write a slash command for linting. You might write a slash command for reviewing a PR. slash commands are prompt templates. So you can pass command line arguments when you run the slash command that will then be interpolated into the prompt template. Couple of UI tips. One, you can use tab to complete files and directories. Cloud code does better the more specific you are. So if you can actually let it know what files or what directories to work with, you'll generally get better results. Hit escape ## UI Tips [05:42] often. I know that I when I started was hesitant to interrupt Claude when I saw it going off path, but you will find your sessions go so much better if you just stop Claude as soon as you see it go in the wrong direction. You can hit escape and ask it to undo its work from the previous turn. And that will help you go back as well. Speaking of undoing Claude's work, I think the biggest failure mode here when working with Cloud Code is you use it to build a project. You get that project to a place where it's working really well, and then it gets overly ambitious, does a bit too much, makes breaking changes, and then you have a hard time rolling them back. ## Version Control [06:21] And the easiest way to mitigate this failure state is to use Cloud Code in conjunction with version control. Ask Cloud Code to commit after every major change. Have Claude Code write your commit messages. there's a good chance they'll be the best commit messages that have ever been submitted to a repository that you own. When working with Claude code, revert more often than what you're used to. Often times, the best way to fix things is just to clear out the conversation history in Claude, revert back to a previous save point, and try again with slightly more specific instructions. Install the GitHub CLI, and it will use this for all of its interactions with GitHub. If for some reason you don't want to install this tool, you can also interact with GitHub via the GitHub MCP server. You can have Cloud Code file PRs for you. You can have Claude Code do code reviews on those PRs. Managing context can certainly be a bit of a challenge when working with Claude Code. You want to always keep an eye on the auto compacting indicator. You always want to know about how long you have until Claude autoco compacts. prematurely ## Managing Context [07:31] compact when you're at natural break points. So if you see you're 35% of the way to autoco compacting and you just finished up a task, you just made a commit, you might just want to go ahead compact right there and start the next task with all of the tokens available to you. Also consider often clearing instead of compacting. Work in such a way that you can use claude code with fresh memory. One way to do this is to tell Claude to use scratchpads to plan its work. Alternative to Scratchpads, you can use GitHub issues. If you are paying per token, then you're going to really want to monitor that uh context window usage and you're going to want to use external memory as much as possible. If you're looking for more robust cost tracking across a team, for instance, one way to achieve that is by using Cloud Code's open telemetry support. So, ## Managing Cost [08:27] for instance, you could hook up Claude Code to Data Dog and produce dashboards that look like this. And for more details on this, you should check out Martin AMP's blog post, which is linked in the description. But in my opinion, the best way to manage your cost is just to upgrade to one of those Claude Max plans at either $100 or $200. I'm on the $100 plan. I spent about $150 worth of Claude Code tokens over the course of about 3 days. If there's a common complaint of Claude Code, it is it's very expensive. So, I was very excited to see Claude Code use bundled in with Claude Max. I don't even know if I got through half of the pro tips that are included in this post. If you want to learn more, uh, check out this excellent post here by Boris and check out some of the other links that I left below in the description.