make claw code 10 times better in one single click. So basically what we're looking at today is this GitHub for Claude code templates. You get a bunch of templates for example for agents, commands, MCPS. You can even create these visual dashboards kind of like a mission control directly for uh Claude Code. Um we're going...winner of that competition actually created this GitHub. So I'm pretty excited to try it out and we'll test it out today. It's called Claude Code Templates on GitHub and it's free to install. Right. So, what we're going to do over here is we're going to grab the GitHub repository. Then
Поиск по транскрипциям
gave you down below that will show you all the custom workflows that Claudebot can do for you now based on this upgraded functionality. " Which brings us to Claude Code. Let's get right into that. So Claude Code, you open it up, you make sure you're upgrading. You'll see Opus 46 there, which is amazing...ways to take advantage of these new upgrades for Claude Code. First thing you want to do is do slash model. You hit enter on that. And one thing you can do now is when you have Opus selected, which is the default, you can use the left and right arrow keys and change the effort level. Right
figures out which ones are relevant. Then it only loads the minimal content it needs. It's efficient. It's fast. And it works across everything. Claude. ai, AI, Claude Code, the API, everywhere. But it gets better. Skills can include executable code. We're not just talking about text responses. We're talking automated scripting, file transformations, real tasks...your rules. Done. Exactly how you want it every single time. Now, let's talk about the risks. Skills can run executable code. That's powerful but dangerous. Claude runs this code in sandboxed environments, but you still need to vet the skills you use, check the code, make sure it's not doing anything sketchy. For enterprise users
fast. Okay, before we start building out our application, let's actually look at a couple different things inside of Claude code that you want to understand while we're doing this. So, first of all, you can always, anytime you're inside of Claude code, you know, just follow the instructions that are there on the screen. So, right...move around with your arrow keys here. All right, you can read through this, and you're going to find all different types of things you can do with Claude code. I won't be able to get into all those in the video because there are dozens, if not 100 plus things inside of there. All right, escape
Okay, so about four hours ago, Claude Code released a brand new feature called Custom Subagents, and we'll be going over that in this video and all the things you can do with it. The docs are also linked down below, so you can either read them or you can watch this entire video because I'll be going...what people have been doing so far. So for example, this person, about one month ago, they said that you can spawn different subagents in parallel by basically telling Claude Code to do that. But in reality, it just spawned different tasks in parallel and had them running in parallel. And other people have been having like an agents
Alright, so after about two weeks, we're back with more Claude Code updates and I'm gonna go through most of the features that were released about an hour ago, of which there are quite a few. Anyways, this video is sponsored by myself and my Claude Code Masterclass. There is a new year sale going on right...there will be a link down below, but it's basically the most comprehensive class on Claude Code that you will find on the internet. Many people have taken the class and gone on to be the best Claude Code users at their company, so you may also find it helpful. Anyways, the first feature that caught
limited research preview, Claude Code added the ability to have it automatically scan your codebase and find security vulnerabilities. So they said that using Claude 3. 5 Sonnet, they found over 500 vulnerabilities in open source codebases. Bugs that had gone undetected for decades despite years of expert review. Now, unfortunately I did not get early access...this is kind of what it looks like behind the scenes. And I imagine they're using a different model in this version of Claude Code compared to the one like on your computer, probably using a model that is less restricted when it comes to like identifying security vulnerabilities, which means I can probably identify more of them compared
very beginning of a brand new project, your CLAUDE. md file may be as small as this, so it just gives a description of what you're making. So Claude Code, when it comes up with a plan, it knows exactly how everything ties back to the bigger picture. Then you have some short commands that may not be inferred...from the code base itself, like using NPM instead of PNPM or bun, for example. And then over time, as you're coding with Claude Code, you may notice that it makes a mistake. And then you add a brand new thing to your file, such as this: "When a library's types are unclear or cause errors, never
right. Now that we've done a little bit of building with Claude Code, we've put together what I would consider to be pretty solid websites with just a few moments of work, let's dive a little bit more into Claude Codes, advanced functionality, and I won't let you guys know that what I'm about...talk about here, probably less than 10% of everybody that currently uses Claude Code understands. So when you unlock what I'm going to be teaching you in this module, you'll know significantly more about Claude Code for one, and then you'll also be able to combine each of these cool different features in fantastic ways that
this up to 80%. And suddenly your business will start growing much faster. Okay. And so you may be wondering why now and what actually is this whole claude code thing? How does it work? So, a little bit of context on that. We have this incredible tool called clawed code. It is a coding tool...certain task. So, it knows how to repeat that for you. And in here, you've got a couple different kind of chats with Claude code open where you're working on different things. So, you usually be working in these different uh sessions. So, one here I might be working on content. One here I'm building
prompt is, then you can run this command over here to use Claude Trace, press enter, and then it will run a version of Claude Code. Basically, that intercepts all the requests going to Anthropic's API, and it shows you the actual requests. So I can open up this file over here. And then if I say hello...calls over here, open up the last message that I just sent, and then type in system over here, and then I can see it says you're Claude Code, Anthropic's official CLI for Claude. And then I can copy the system prompt over here and go to a new doc. And then I can actually see the system
like our own employees, right? We basically have two employees working for us here instead of us having to handhold an AI agent like you would in Claude Code. I now have employees working for me. And now I can come over here. I have the codeex plugin inside a cursor. Now I can get to work...actually very easy to integrate all the code from the different AI agents. The main AI agent inside of the codeex plugin is working just like you would a Claude code. All the code changing here as you can see on the right hand side. In a second, it'll be complete. It actually breaks it down to task
have an idea you can just build it. And it's a lot more about the idea now than it is about the details, because just like Claude Code, you can rewrite the code over and over. And, you know, Claude Code itself, we rewrite all the time. And I think this is just something that coding agents enable
this case here. And in the end, it is always all about constructing the right context and giving the agent the right information. So skills, just as with Claude Code, still matter a lot when working with Codex. And I found Codex to be very efficient at, when it comes to activating the right skills at the right time...really happy with its behavior And then there are these points here, which I also already mentioned in the Claude Code video, and they have not changed, but they're very important to me. You should absolutely embrace these agentic engineering tools. They are the future of development. I'm convinced about that. And you can be super efficient when
There's been a bunch of new updates to Claude code and I think at this point I've stopped using Open Claw most of the time and I want to talk you through why that is um and why I think Claude is actually better at this point versus open claw. I'll still be doing updates...skills that Claude has received um with the latest features etc. There's actually 10 of them that I'll walk you through. These updates actually make Claude Code and Claude in general better and more powerful than something like OpenClaw in general, right? You can see, for example, yesterday they actually launched inside Claude Code the new schedule task
minute, and bakes it right in without all the fluff, keeping the agent side of things at the forefront. And I think what happened is when Claude Code brought the AI in the terminal to the IDE, it completely eliminated that context switching that we all hate. And that's what warp is recognized as well. No need to leave...building something and then you make all the little tweaks you need by hand. So, that's a huge improvement and I think brings it to par with where Claude code is, but with its own particular flavor and perspective. You don't integrate one with the other. Warp actually just has it all in one package, one single environment
main session. So if you were to do everything inside of the main session, it would fill up with a bunch of random noise and like output that Claude Code did not need. And that would lead you to use your context window of the main session faster. And we know that as a context window fills up, then Claude...main session. And whilst in this example, it happens at the start of the conversation, it can also happen in the middle of the conversation. So your main Claude Code session may be doing some stuff. It then decides it needs to call a subagent, and then a summary of everything that the main session has done so far alongside
deployment directly from claude code in the future. So again on the left hand side you're going to go here to generate token. For the token name I'll just give it a never expires and I'm just going to call this claude code and then generate. Okay. Now I'm going to copy this token. Make sure...results. " All right, so I asked it this, but for some reason it wasn't able to find the MCP server. And I believe it's due to how claude code is kind of working directly inside of cursor. Now I'm just going to quickly show you that if we go and we go to the settings here
long as you like. Now, let me quickly show you the pricing. So, the $20 monthly plan, if you pay monthly, that gets you Claude Co-work and Claude Code, too. But again, it's going to be limited on how much usage you get out of it. I actually quickly hit my limit here. This is the plan...dependent on usage. Starts at $100, but right now I believe I'm paying $200 for that. And they also have T-Mn enterprise plan that also unlocks Claude code and claude co-work. Now, hopefully it's really obvious that Claude Co-work is one of the closest updates in AI that I've ever tested that feels like
really interesting watching it go. Okay, so it seems that both of them are done. Codex CLI did use more tokens as well, it used 130,000 tokens, whereas Claude Code used 72,000 tokens. So I expected more tokens, it probably got more context from online or something, so I'm hoping the solution will be better...with a slightly different like, padding, color scheme, and design. But yeah, as for the underlying logic, it works really well. And now this is the one that Claude Code gave us, so we can go to MCP servers. It did copy the design over, like the design is very similar to the one in models over here