Transcript: Give Me 50 Minutes, I'll Give You 1000+ Hours Of Claude Code Knowledge (2026 Guide)

Watch video

Chase AI51:28Transcript ✅Added Aug 12, 1:51 pm GMT+8

Source video ID: U6k4MeVks_Y

Transcript

0:00 — This video is going to replace over 1,000 hours of trial and error inside of Claude Code and tell you what you need to focus on whether you’re at the beginning stage, intermediate stage, or pro. I’m going to cover everything from the desktop app to how to prompt, MCPs, CLI, and even more advanced topics like graph engineering and creating your own agentic OS. And by the end, you’re going to have a complete road map on how to master the most powerful AI tool in the game today. There is a ton to cover, so let’s get started. Now the first thing we need to talk about in the beginner section is where
0:31 — do we run cloud code because this can actually be a confusing question. You technically can run cloud code from the web app in the cloud. We also have the desktop version of cloud code. And of course we have the terminal. Which of these three should you be using? Especially if you’re just getting started. Well, if you’d asked me this question a few months ago, I would have pushed you to the terminal or something like VS Code with the cloud code extension. But these days, the desktop app has gotten a lot better. And if you are someone who is not coming from a technical background, this is all brand
1:01 — new to you. I would suggest using the Claude Code desktop app. Yes, the terminal will always be a powerful option for those of you who are more technically inclined. But if that’s not you, don’t feel pressured into using it if you’re not comfortable. You really aren’t missing out these days. And the desktop app is getting features that just aren’t available inside of the terminal. things like voice mode, things like browser automations, and there’s a lot of quality of life things when you work inside the desktop app, like inline artifacts, and generally a better user
1:31 — experience for those of you who have never touched the terminal before. Not to mention, we can also use the terminal when we’re inside the desktop app. So, it’s not like it’s an eitheror thing. Now, to download and install the Cloud Code desktop app is super easy. Just search for Claw Desktop App. You’ll hit the first link, and then you’re going to download the installer and run it. So now you have the desktop app installed and let’s do a speedrun through this thing and what you need to actually care about. Now in terms of settings, you come over here to the left and hit customize. Let’s then go to the general tab. And I want you to take a look at instructions for Claude. Take a look at mine. It’s blank. Yours probably should
2:02 — be blank, too. These are global instructions. By global, this means it applies to everything, all your projects and all of your prompts. So if I put something in there like, I want all of your responses to me to rhyme. Well, guess what’s going to happen? Every single response is going to rhyme. stupid example, but the point is, do you actually have something that’s relevant enough to every single chat you’re ever going to have with Claude that we should spell it out here? Maybe, but that’s a very high bar. And if what you’re thinking about adding here doesn’t meet that bar, I would leave it blank. Next,
2:32 — go to capabilities. Make sure tool access mode is load tools when needed, and turn everything else on this page on. In the cloud code tab, I would keep everything in general turned on. It has some settings that are purely personal preference. And in terms of local sessions, I would turn all of these on as well. Now, the exception is when we come down to pull requests. If you don’t know what a pull request is, just leave this off. It’s a little more advanced. And for Claude in Chrome, which is a Google Chrome extension where Claude can control things in your browser. I turn
3:03 — this on as well, but understand you need to download the Google extension if you want to do that. Now, we’ll go into more detail when it comes to skills, connectors, and plugins, and memory later, so don’t worry about that just yet. Now over here we have artifacts. You don’t really need to worry about artifacts honestly inside of Cloud Code. Routines have to do with automations we run. And we’ll go deeper into automations later. And over here this is just previous chats we’ve had with Cloud Code. Now if you go ahead and hit new, it should bring up a page that looks like this. And let’s break down what we’re looking at. So obviously right
3:33 — here we have our little chat window. And then we have these four things up here. Local, something that says a th00and main, work tree, and then a little plus button. Local is just you telling Claude where it’s actually going to be running. If you don’t know what any of these things are, cloud, remote control, WSL, SSH, you should be on local 99.99% of the time. Everything else here besides WSL is all about being able to use Claude when you’re away from your computer. So, if you’re in the beginner stages, that’s not going to be you. Next, right here where it says 10,00
4:03 — that is just the folder I am working in. So if I click on this, I can open a new folder and I can pick any one inside of cloud well inside of my computer that I want to work within. So you could create a folder on your desktop that just says cloud code projects and that’s what you work in. Everything you do here inside of cloud code will live in that folder. So you just have to pick one. Next we have main and workree. This has to do with git. A little bit more of an advanced topic. So if you’re a beginner to oversimplify this completely, git just has to do with saving your work.
4:33 — So, if you don’t know what git is, I’m not going to turn this into a git lesson. Just keep it on main and don’t check workree. And then we have this plus button. This allows you to add an additional folder so your work essentially gets copied in two places. Next, we have permissions. If I click on auto, I will see five different modes. What are these modes? This has to do with us telling cloud code what it has permission to do with or without our consent. So, on one end of the spectrum, we have manual. This means it’s going to ask you all the time, can I do this? Can I do that? Can I edit things? On the other end of the spectrum, we have bypass permissions. It can do whatever
5:04 — it wants. Download things, install things, delete things, edit things. Kind of scary. Now, in between those, we have auto, which is essentially bypass permissions, except there’s a classifier that takes a look at the commands that Cloud Code is running and decides if they’re dangerous or not. And if they’re dangerous, it will stop them. So, this is the default for a reason. You should be sitting here all the time. The only other one we’re going to play around with is plan, and we’ll go into that more later. This plus button lets you add things. So you can add screenshots, that sort of thing. And then you have a microphone. And then over here on the right, we have the model, we have the
5:35 — effort level, and then we have our context window. What model should we be using? Well, it depends on what plan you’re on. If you’re on the $20 a month plan, you really aren’t going to be able to use Fable. It’s just going to burn too much usage, and so you’re going to have to sit on Opus. If you are on a max plan, 5x, 20x, so $100, 200 bucks a month. I would suggest working in Fable most of the time. It is the best model by far. Now, the problem is the usage though because if we click this little thing over here, you know, you see context window. We’ll talk about that more later. We have a number of different limits. We have a 5 hour
6:06 — limit, a weekly limit, and a fable limit. So, only half of our usage per week can be dedicated to Fable. So, we don’t want to burn too much Fable too early. So, what we need to really think about is our effort level, which can be anywhere from low all the way to ultra code. The more it thinks, the better it performs, but it’s not linear. If I go from extra high to ultra code, does that mean I’m going to get like a 10x increase in the performance? No. It
6:36 — might be like a 1% increase, but you might pay 5x more. In fact, most of the problems you’re trying to solve, especially if you’re more in the beginner stage, don’t require anything beyond medium. In fact, you could probably get away with a low and just do fine. Me personally, I sit on Fable medium most of the time unless I’m dealing with something rather complex or I’m about to hit a usage reset and I know I can just burn it. So, for today, we’re going to sit on Fable 5 medium because it’s a nice happy middle ground. So, lower effort means less usage. Doesn’t perform quite as well, but generally it does more than enough. Now,
7:07 — let’s talk about prompting. Anytime we are going to start prompting some new project, I highly, highly suggest you go into plan mode. Now, why is that? Well, plan mode is going to allow Claude and I to have a conversation to make sure we are on the same page before it goes and executes something. And more importantly, not only are we going to have a conversation, it’s going to ask you questions because many times, and this is very true if you are coming a from a nontechnical background and b if you’re trying to do some sort of project that is not in your typical, you know,
7:38 — domain of expertise, the issue you’re going to run into when using AI is that there’s just so much that you don’t know. And there is so much that you don’t know that you don’t know. These unknown unknowns are a real problem. And the only way you’re going to be able to figure those out is by having clawed code pretty much bring them up for you to shine a light in these dark spaces that you didn’t even know exist. And plan mode is the simplest way to solve that problem because it’s going to force it to ask us questions. So anytime you’re like have a fuzzy idea of I’m on step A. I want to get to step Z. I don’t know what to do. We’re going to go into
8:08 — plan mode. Now, when I prompt clawed code, this used to be a big thing kind of back in the day, the last like year or two, and this has kind of gone away a little bit, but you still have people out there who think there’s some like magic prompt you need to do. It needs to go in this specific format where like here’s the goal and here’s the context and here’s how I want you to act. You don’t need to do that. What you need to do is you need to buy yourself a microphone. You need to turn the microphone on and you just need to give Claude Code a stream of consciousness. So, for this plan, for this website we’re going to build, we’re going to say it’s just a website for this fake AI
8:39 — analytics company, and we’ll call it Lighthouse. And so, that’s all I’m going to do. And I’m just going to ramble like there there’s no plan here. It’s going to sound like this. So, I want to create a website for a fake AI analytics company called Lighthouse. I don’t really know what I want to be on the website. Although I know I want sort of the call to action at the end for them to book a call with us. So that’s going to be kind of the call to action.
9:09 — I think in terms of who the target audience is. Let’s say it’s going for like small startups. So other than that, I don’t really know what I’m missing. Just go ahead and ask me whatever questions you think are relevant that I haven’t thought of. Now what I want you to pay attention to is that last bit where I said pretty much just start asking me questions and things I haven’t thought about. Now, since it’s in plan mode, it’s kind of already going to do that. But anytime you prompt Claude, you can always add that at the end of your prompt, like what am I not thinking about? What sort of questions do you have for me? You know, and this is going
9:40 — to get you again in sort of that back and forth with cloud code because at any point it can bring up something like this where it asks you questions. So, what’s the purpose of this fake site? Well, we’ll say it’s for design/dev practice. What does Lighthouse actually do? We’re going to say product analytics and AI insights. How big should it be? Let’s do a landing page like that. And then what does vibe should the design have? We’ll go with clean light SAS. Next, it asks us what
10:11 — text stack should we want the site to be built with? Plain HTML, CSS/JS, next.js plus tailwind, or ash plus tailwind. Do you know what any of those are? Genuinely, do you have any idea what those are? Do you even know what a tech stack is? If your answer is no, then what should we do? Should we just go with the recommended? Yes, but no. So, here’s a problem a lot of people run into is they’re just going to be like, I don’t even care what a text stack is. Sure, plain HTML. Go recommended. They just click this. And the thing is, these models are so good is that you’re still going to get a
10:41 — pretty good output. The problem is is when you repeat that behavior of not understanding the question and just hitting recommended again and again and again and again. The main problem is there’s no differentiation between you and any guy on the street who I could put in front of your computer and have them do the exact same thing. Like where is any moat whatsoever for what you do? Like you’re very replaceable. But even more importantly, you’re not learning anything. And as good as these models are, eventually you’re going to have a pet project that’s very unique that maybe Claude doesn’t know the best way to attack it. If you have spent all of
11:11 — your time learning cloud code by never actually learning anything and just hitting recommended, recommended, recommended, you’re going to have zero idea of what’s actually happening. Now, you don’t ever need to learn code again, but you do need to begin to learn AI software engineering fundamentals, like big picture stuff, like how these building blocks come together. And the only way you’re going to do that is when you hit questions like this, you don’t just hit recommended. You say something like this. Can you explain what a tech stack actually is? I don’t really know.
11:41 — I don’t really understand these options as well. So, can you just give me a quick breakdown of what I’m looking at? And that’s it. You’re just going to tell it to explain the question a little more detail. And if you do that over and over and over again for weeks, months, and years, you’re eventually going to build an actual foundation. You’re not going to be this like caricature of a vibe coder. This is extremely important. And this is the sort of mentality you have to take with prompting and talking with claude code if you actually want to like learn anything because it’s going to make you better. So we’ll do pricing product features. Sure. What should the
12:13 — book uh call CTA work? We’ll do a fake booking form and it’s off to the races. And so we can see here it broke down the what is a text stack question and then goes into a little more detail about what everything is. Again, you don’t need to become an expert at all these things, but as you do this again and again and you kind of go down the rabbit holes when they make sense for you, you you generally are going to start like putting the pieces together. Like this isn’t that complicated.
12:44 — Like coding is difficult. No one’s asking you to become like an software engineer in that sense. But you can learn what a tech is. You can learn what these different languages sort of are and like what we should do in each sort of use case. Now once it proposes the plan you will see it populate over here on the right hand side. Now what’s cool about plan mode inside the desktop app is when it brings this up at any time I can sort of select something. So let’s say the audience let’s say I didn’t just want small startups. Let’s say I was also looking for add I wanted to add
13:15 — medium sized companies as well. So if I do that and I hit comments it now starts adding sort of like comments here. So I could add more and more comments and go down the line and kind of make these like almost little notes to the plan it’s created. I also have the ability to add any prompt I want and then I can at any point say hey let’s go ahead and revise this. So now it’s added mediumsiz companies to the audience as well. And once I like the plan I can either hit accept or accept with auto mode. So make
13:45 — sure you hit accept and auto mode or else it’s going to start running it inside of manual mode and we don’t want to do that. So we’ll do accept and auto mode. And now it has gone ahead and built the web page for us. And it’s at this point we are going to move into the intermediate sort of skills and tips and tricks when it comes to claude code. But before we do that, a quick word from today’s sponsor, me. So just yesterday, I released a completely updated version of my Claude code masterass. We go way deeper on all the topics we sort of touch on at a surface level here in today’s video, and it is a perfect place
14:15 — for you if you’re someone who is non-technical and really just wants to learn how to master this amazing tool. So, if you want to get your hands on it, you can find it inside of Chase AI Plus. There’s a link in the pin comment. So, let’s build our website. But what I actually want to focus now is on this guy down here, this little circle. You remember this that showed our usage? Well, it also shows our context window now. And if we click on our context window, it gives us a very specific breakdown of what is actually filling it up. Now, the context window is a very
14:45 — important metric we always need to keep track of for a few reasons. Now the first reason has to do with performance. But to understand this, you need to understand tokens and context. So to keep this oversimplified, every single word you send to claude code and every single word you get back is considered a token. A token is the currency of large language models. And the context window is the budget. So we have a budget of 1 million tokens we can essentially spend in each session. And so far we have spent 156k.
15:15 — Now that sounds great. We’ve only used 15%. I can use another 844,000. Well, sort of. The issue is as this context window fills up, in fact, the performance of claude code gets worse. Think of it as it just having too much stuff in its brain. If we are sitting at 800,000 tokens over here out of 1 million and I ask it questions about what’s been going on in the last 800,000 tokens, it’s going to struggle. Especially if we’re asking it questions that sort of happened in the middle. So
15:46 — because of this, we always want to keep an eye on our context window because we don’t want the performance to decrease. And this decrease is somewhat linear and there isn’t an exact science to it. So rule of thumb as we sort of hit like 30%, 40%, certainly 50%, 500k tokens, you definitely want to sit there and ask, do I need to continue the session? And really for me, that’s at like 30%. And while you can see here the context window does get filled up with things besides our messages, including things like system tools and skills, the
16:16 — messages are the big thing. Let’s say you hit 30% of your context window or even 50%. What are your options? Well, we really have one option and that’s just to start a new chat. Okay, we’re just going to start a new chat. Now, there’s a couple ways you can do this. We can do commands like forward slash clear. If I do forward slashclear, this is going to get rid of everything and we’re going to start completely fresh, brand new context window, and we’re
16:46 — going to have top performance. Now, the other option is to do slashcompact. If I do slashcompact, what cloud code’s going to do is it’s going to take a look at the entire conversation history we’ve had. It’s going to create a new summary, and then it’s going to start a new chat with that summary. Now, your other option is just to go over here and hit the plus button, and this will also start a new chat inside of that same folder, and then I can reference the old chat at any time. Now the scary thing is especially if you’ve come from primarily the web app where the conversation you have is sort of all you have and if you get rid of the conversation it forgets
17:17 — everything. Remember what are we doing here? We are working inside a specific folder. We’re inside folder 1000 creating our website. So if I get rid of this entire conversation it can still take a look at the files at all the code we wrote and understand what’s happening. So starting a new chat isn’t really starting from zero. And in that case, you really have nothing to fear if you’re filling up the context window and you’re scared of starting over. It’s okay. Just start over and worst case, just have it create that summary with /compact. But at this point, we’re only at 16%. So, we are okay. Now, let’s talk
17:48 — a little bit about this website. This website is ugly. This website is actually pretty lame and generic. And by the way, I’m looking at this inside the browser pane inside of Claude Code desktop app. So, I can actually do a lot of stuff here. Like, if I open this up, I can select certain things. I can put comments on these just like in plan mode, which then go into a prompt. I can actually annotate things and say like, “Hey, trash, you know, and it will also add it
18:18 — to a comment.” So, really easy to sort of do like micro edits here, um, if you want to. But big thing we need to solve for is just like this website. Why does this look so terrible? Well, this looks pretty awful because a we didn’t give it enough context. for what we want it to look like. We didn’t give it any sort of inspiration. We didn’t give it screenshots. All we said was we wanted like a clean SAS product. And one of the themes of sort of the intermediate section is context engineering. Your ability to give cloud code access not just to sort of your thoughts and your
18:48 — vision, but for it to external tools and skills so that it can do a better job. So sort of a two-part process here for solving this problem. This is where we’re going to dive into the idea of skills. Skills are probably the most important thing you need to understand and master when it comes to leveling up your Claude code performance. Now, skills at their most reductive state, and it gets a little bit more complicated, are simply prompts that tell Claude Code to do a specific thing in a specific way. For example, there are a ton of front-end design skills
19:18 — that are all about making better websites that are essentially just prompts that tell Cloud Code, hey, uh, when you’re creating this website, avoid certain gradients, avoid things that look like AI slop, do this, do that. Right? It’s just giving it specific instructions. That’s all skills are. So, how do we actually get skills? Well, we can find them inside of the Claude app. So, if I go to customize, move this over here, we go down to skills, and you can see some of the skills that I have right here. We also have plugins, which also can kind of be skills. It’s sort of a gray area, especially in the
19:49 — app. When we talk about skills versus plugins, you can kind of think of them as the same thing. Plugins can include multiple skills, but it’s kind of arbitrary. So, for example, if I go to plugins and I go to browse, first of all, what am I going to see? I’m going to see the anthropic official plugins. This includes the front-end design plugin, which is simply the front-end design skill. So, if I install this, mine’s already installed. It will then add the front of design skill to claw code. And this is the actual prompt for the exact front of design skill. This is an official skill. Like you can just go see this. This is on the official claude
20:20 — code GitHub. If I copy this whole thing and I go back inside here and I paste this into the prompt, that’s just like I’m using the skill. But obviously you wouldn’t do that every single time you want to do something related to front design. So instead we simply add the skill like I showed you in the UI and we just do frontend design. Right? I can do forward slash. And now this is invoked. Now that’s the same as if I just copied and pasted that whole thing. Now I don’t only have to do the forward slash. I could just use natural language and say like use front-end design skill and it’s
20:51 — smart enough to actually know it needs to call that. Now the confusion can happen is if you have multiple front-end design related skills and if you do that then you if you just say like hey I’m building a website then it might not know which one to pick. So if you have multiple skills that all kind of do the same thing you need to nudge Claude in the right direction. Now, the most important skill you can actually add right now, and this goes beyond web design, is if you go back to plugins and you go to browse and you go to skill creator. This is the most important skill you can add because this is a skill that allows you to create other skills. It includes things like
21:23 — measuring skill performance. It runs tests. It runs evals. It does benchmarks. We’ll talk about this more in a little bit, but obviously as we look through here, there isn’t that many skills to choose from. And we all know there’s a billion in one skills floating out there in the world. And so where you will normally find skills is GitHub kind of like you see here. So let’s say for example I wanted to use another front-end design related skill. And let’s say I was looking for the impeccable skill and I found it on GitHub. Well, how do I actually install this thing? So it can will tell you in the description how to do it step by step. But often times it’s kind of a pain in the butt. All you need to do is
21:54 — when you find a skill you like is just copy the URL from GitHub, go inside of Clawed Code, paste that skill in there, and then say something like add this skill. From there, it will literally add the skill to your repertoire. And from there, you just invoke it like I showed you. Now, you can get pretty advanced when it comes to skills, especially when you have the skill creator skill. For example, let’s say we finish out this video. I add a bunch of new stuff to this web page. I could at the end of doing all this of like prompting it,
22:25 — creating the website, doing the additions, adding whatever I want. I could then do something like, let’s use the skill creator skill and then say, take a look at the entire message history of the session. Take a look at everything we’ve done today and turn that into a skill. So, if there’s things that you do over and over and over again, you can turn those into skills. And later, I’ll even show you how to then turn those skills into automations. So skills are very powerful because they allow you to codify things that AI does. You know, one of the issues with AI is that it’s somewhat non-deterministic, right? If you ask it to do something 10
22:57 — times, it might do it 10 different ways. It’s not deterministic. However, skills allow us to be somewhat deterministic and give us more control how Cloud does things. So that’s why they’re so important. Now, like I mentioned before, beyond skills, we also need to add more context here. So what I’m going to do is I’m just going to search for some screenshots that I can add to this website to make it look a bit better. So, I went on Pinterest and I found this image when I put in SAS landing page. Thought it looked kind of cool. So, what we’re going to do is we’re just going to drop this screenshot into here. And we’re going to say, “Use the front-end design skill to redesign this web page.”
23:29 — And in fact, I want you to do three versions of it and show me all three versions inside the browser pane that I can choose from. All of them should kind of be in the style, but I want it to be divergent enough that I can kind of see some differences. So, it did exactly what we asked. If we look at this here, we can now see the three different versions of the website. You can see what a departure these are from what we were just looking at. And that was with a single skill, which was just the generic front of design skill from Anthropic, and a pretty basic prompt
23:59 — alongside the screenshot. So, I can take a look at this one. V1 full size. This one looks pretty cool. We have this V2. Um, not a huge fan of the colors, although the radar looks kind of cool. And then lastly, we have V3, which kind of looks like your typical AI slop to be totally honest. Um, I really like V1. I think this is something that looks pretty cool. So, what we’re going to do is we’re going to say, hey, we’re just going to go ahead with V1. What that’s meant to demonstrate is really the power
24:29 — of just injecting context. Literally, one screenshot, one skill, infinitely different end result. So now let’s talk about how we can supercharge cloud code by bringing in outside tools by connecting outside applications to cloud code itself so that claude can control them and we never even have to really leave the app. Now there’s really three different ways to do this. The first one is by going into customize and heading to connectors. Some of these are super easy to connect and you might have done already done it already. That’s something like Gmail or Google calendar
25:00 — or Google drive. This allows Claude Code to talk with these applications, control these applications, usually with some guardrails, and it just happens via prompting. So, if I tell Cloud Code, hey, go read my Gmail, it does that because it’s been connected. Now, most of the big apps out there have some sort of connector. So, if you just go to add and you browse connectors, chances are you will find what you are looking for. The second way is via plugins. And like I said, there’s very much like a fuzzy line between all these things. similar to connectors. If it’s big, if it’s
25:30 — popular, there’s probably a plugin for it. So, if I go to browse, you’re first going to see a bunch of anthropic plugins, which really are just skills. If I go to partners, though, I can find something like GitHub or Superbase. I simply click on it, and you’ll see what’s going on under the hood. In this case, this is the GitHub MCP. Just like with connectors, if I add some sort of plugin, it’s going to allow me to talk to control some sort of outside application via cloud code. But there is a third thing that is not an MCP or
26:00 — plugin or connector and those are things like CLIs. So we have right here the GitHub CLI. So we have a GitHub CLI and we have a GitHub MCP. What is the difference? Well, the difference is a little bit technical in terms of the practical application. What you are going to care about in many cases there’s not a huge difference. In general, the CLI tends to give you more functionality than an MCP. And often times the CLIs also include skills. So when you are dealing with some sort of outside application, aka you’re working
26:31 — inside of cloud code and you need to talk to something else, you need to figure out if you can add it through one of those three ways, connectors, plugins or CLI. Like you saw connectors and plugins, we can do it through here via the CLI. It’s also as simple as simply telling Cloud Code to add the CLI. So like the GitHub CLI, there’s an actual command if you’re inside the terminal, but I could just like copy the URL. I can go inside of Claw Code and say like, “Hey, here’s the Cal CLI for GitHub. Go ahead
27:01 — and add this CLI.” And it’s going to do exactly that. From there, you’ve pretty much given Claude code. Think of it almost as a skill to call on that CLI and have it do whatever that CLI gives it the ability to do. For GitHub, that means I can create a repository. I can upload all this code we just created to that repository. I can edit the repository. And if we take it a step further and we think of something like Verscell and Verscell, if you are unaware, is a web application that allows us to host our actual website. So
27:32 — we just created a website and we want to put it on the web and have a true URL. Verscell will allow us to do that. Well, instead of going to a dashboard like this inside of Verscell and handling all ourselves, why don’t I just search for the Verscell CLI? Oh, look at that. forcell has a CLI as well. And so what we could do with just those two applications is I can take a look to see if there’s a GitHub connector. There is. I can take a look to see if there’s a GitHub plugin. There is. I can take a look to see if there’s a GitHub CLI. There is. Add any one of those, add it with Versell, and I now have a pipeline
28:05 — where I can take the website I’ve created. I can create the repo for it inside of GitHub and then automatically connect that to Verscell. I’ve essentially created an entire deployment pipeline from cloud code that I just have to talk to in plain language and it’ll do everything for me. And the big takeaway here is that anytime you are working with anything outside of cloud, you need to ask yourself, can I actually just control it with Claude? Because chances are Claude can actually control it better than you can, especially if you aren’t intimately familiar with how that application works. Whether it’s a CLI, an MCP, or a connector, it doesn’t
28:36 — really matter. You just need to pick one. And you also don’t even have to like go out and find it. So like for example, if I asked Claude code a question like this. Hey, so I’m thinking about deploying this website. I’ve heard of things like GitHub and Verscell. I’m not super familiar with those. I’ve also heard there’s like CLIs or perhaps MCPs we could use. So do you think you can go look and see if that would make sense for hosting our website? If they do, can you add those CLIs if we need them? And
29:06 — then once they’re added, can you go ahead and just set up that deployment pipeline and get it all properly connected? So a prompt like that where you’re just saying like, “Hey, I heard there’s some tools out there. Maybe they have CLIs. If they have them, add them. Run them.” That’s all you got to do, right? You don’t even really have to be a pro. When in doubt, just ask Cloud Code what the best practice is. And if it has a CLI, it’ll go ahead and search for it, right? So if I run that, and by the way, I already have these installed, so it’s probably going to come back and say, “Hey, I already installed it.” All you have to do if you’ve never used them before is create an account and then we’ll walk you through the setup. And as
29:37 — these agent coding harnesses like Claude Code only become more and more ubiquitous, you’re going to see pretty much every app out there come out with some version of a CLI, a connector, an MCP, which means again Claude runs the show. And so with that prompt, obviously it’s going to walk you through how to log in if you haven’t done that before. It created the GitHub repo. So all the code from our website is essentially living in the cloud. And it also set up the Verscell connection. So I now have a live URL I can go to and we can also see it right here. So this has a real URL lighthouse site to versel app and I
30:08 — could share this with anybody and again did I have to go into GitHub? No. Did I have to go into Verscell? No. All control from cloud code. And because it is a GitHub and Verscell connection any changes I make to the website inside of Cloud Code here, if I want those to be reflected on the live website, I just tell it and it does it. So this deployment pipeline is just one example of bringing outside tools into the cloud code fold. Now it’s time to move on to some more advanced topics. We’re going to touch on things like automations. We’re going to talk about how we can best approach long
30:39 — horizon tasks. We’ll do that by discussing things like /goal loop engineering, graph engineering. We’re going to talk a little bit about model routing, how we can bring in other models like codecs and the GPT models into our workflow. And then we’ll finish it all off by talking about some more custom type harnesses and UIs we can layer on top of cloud code. Whether that’s something like this or going with an obsidian command center type approach. So let’s begin by talking about long horizon tasks and loop engineering and this kind of bleeds into things like automations. So when we talk about long horizon task, we’re talking
31:10 — about things like loop engineering and graph engineering. What are we really saying? What we’re saying is we have some sort of task, some sort of goal we need to complete, but that this might be something we need cloud code to do an infinite amount of times. This might be something that runs every single day. And ideally, it is something that not only is going to run every single day, but something that we want to be self-improving. So for these long horizon tasks, there’s really three parts. We are going to have a trigger, a task, and then some sort of success criteria. Now, not all long horizon
31:42 — tasks are loops. This could simply be some sort of task that you think is going to take 2 hours, 4 hours, 12 hours, days for Claude Code to complete, and you don’t want it to just stop every single time it fills up its context window. You want it to keep co going and going and going until it completes that task. There is a built-in command inside of Claude Code that does this. It is called slashgoal. Now, slash goal is perfect if you have a complex project that you want Claude Code to complete and you don’t want to babysit it throughout the entire process. However, there is one specific thing you need and
32:12 — that has to do with success. You need to be able to define success because slashgoal isn’t the only thing you pass. You also have to pass it a prompt. And that prompt needs to be what the success criteria is. Okay? Like you have a goal for claude code to do. What do you want it to do? It’s not enough for you to explain, hey, I want you to do X, Y, and Z. No, like what is the end state? Because what’s going to happen when we run forward slashgoal is it’s going to attempt to complete that goal with code
32:44 — or whatever it’s going to do. It’s going to run its first iteration. It’s then going to compare that iteration to the success criteria you defined. If it meets the success criteria, cool. It’s all done. If not, it’s going to pull up a second session, run it again. It’s going to go ahead and check the success criteria. Did it work? No. Then it’s going to run it again. Now, each time it runs, it’s going to take a look at its previous iterations to see, hey, what worked, what didn’t, but it’s going to continue sort of this internal loop until it reaches its goal. And this is
33:15 — extremely powerful. It’s kind of similar to Ralph loops, if you know what that is. And so, it’s important that our success criteria is as objective as possible. If I’m just saying, hey, the goal is to create a cool website that looks cool, right? That looks neat. How does it know what cool is? How does it know what neat is? How can it actually take a look at the end of every single run and say, “Did I complete this or not?” So, the more subjective your criteria is, the sort of worse it’s going to get. Not even necessarily the
33:45 — worst it’s going to get, but the less likely that it’s going to meet your needs. But with all that being said, goal is definitely a form of loop engineering, but it’s one that pretty much has a definite end. We aren’t going to do forward/goal and expect this to run for eternity. But there are things we may want to do where we do want it to run for eternity. We still want it to act in the same way as goal. We want it to trigger on demand or perhaps on a schedule. We have a task we want it to complete. We have success criteria and
34:15 — we want it to also be self-improving because again this is self-improving in an aspect because it’s always taking a look at its outputs and comparing it against some sort of success criteria. So what happens if we want to create something that does sort of loop forever, right? Perhaps it’s an automation that runs every single day that we want to continuously approve upon itself. Now we will continue with these three steps, but next we’re going to have to add some sort of logging phase. So now let’s take a look at a custom loop or an example. We have cloud
34:45 — code right here and what we want cloud code to do every single day is we want it to create some sort of morning report for us, some sort of morning brief. I want to go out on the web. I want it to find AI news for me. I also wanted to check my Gmail and at the end give me some sort of document. So cloud code is going to take a look at YouTube. It’s going to take a look at Twitter, Reddit in my Gmail. It’s going to grab all that information, scrape all that information, consolidate it, synthesize it, and give it to me in a report. And I want it to do it every single day. So
35:16 — how could we bring loop engineering fundamentals into this? Well, remember we just have to set up these four things. So, what is the trigger going to be? Well, let’s say the trigger is every day it’s going to run at 7 a.m. What is the task? Well, the task is just what I described. Scrape these websites, consolidate, create a report. So, that’s set up. Well, what is the success criteria? Well, this is this is where it gets difficult, correct? Because this is something that’s kind of objective, like what makes a good report a good report.
35:46 — We could add some sort of subjective material there like hey every single report must have at least five videos from YouTube and five Twitter posts and five Reddit post and you must call out XY and ZZ on the Gmail. So there are some things we can do here but it’s not as simple saying like hey you’re doing a loop on some Python application and your goal is to get it to this particular speed. And then lastly we have logging. So every single one of these reports we could put it in some sort of database, right? That way clawed code can always take a look at its past work and compare
36:17 — its upcoming work to things we have already done. And this in essence from a theoretical perspective is loop engineering. Now what does that actually look like in terms of a practical sense inside of cloud code? Well, step number one would be creating some sort of skill. Because what did I just describe here? Well, I just described a skill. We can create a skill that cloud code runs on command or on a trigger where it does all this stuff where it scrapes information and turns it into report and we can include in that skill it’s sending all that information to a
36:47 — particular database. So step one to do sort of loop engineering for real would be to invoke the skill creator skill. Remember I showed you how to do that earlier. And then you would just describe the skill as I just did. From there you would then run the skill manually over and over and over until you got it to a pretty good spot. Once you are happy with that skill, you would make sure you add a language about, hey, I want this to be logged in a database and every time we run the skill, I want you to take a look at the previous
37:17 — iterations and see if we can do better. That’s sort of where the self-improving aspect comes in. Ideally, you would be able to score every single one of your previous reports so it has some sort of objective measure it can base its outputs on. But simply, you’re first going to turn it into a skill. Once you’ve turned it into a skill, all you have to do now is turn it into an automation that runs all the time. And this is really simple inside of Cloud Code because we’re just going to turn it into a routine. So, if I come here to the left and hit routines and I go to new routine and I go to local,
37:48 — guess what I’m going to do? I’m just going to tell this to run that skill every day at a particular time. So, this would be like loop skill. We are running the loop skill and then the instructions would literally just be run the slash loop skill. Okay, when you’re doing this and you’re creating this inside of cloud code with the skill creator, you would also say like, hey, I’m trying to do this to so it’s like self-improving. I want to do the database loop engineering
38:19 — fundamentals, etc., etc. The great thing about the skill creator skill inside of cloud code is that it’s going to do all the heavy lifting for you. it understands sort of like the goal here. And then from there, you just schedule it, right? Ideally, it’s probably just going to be something that’s daily, but you can do it hourly, weekday, custom, whatever you want. And that is the practical application of loop engineering. And that’s pretty much all you need to know because beyond loop engineering, we then start talking about graph engineering. And graph engineering can get a little more complicated. But what’s really happening is remember
38:51 — before when we looked at this, we had all this loop engineering stuff going on. Let’s actually just let’s undo all this. We got all this loop engineering stuff going on, right? Like, hey, this one guy is scraping everything. It’s looking it’s creating the PDF and then it’s being graded. Well, what if we did all that loop of trigger of task success criteria and then grading it? You know, logging the information. What if we did that at every step of the journey? So, we had one agent that does this with scraping YouTube. So, it has a trigger, it grabs YouTube, it scrapes the data,
39:23 — and then it judges how well it scraped it by looking at its past iterations. It self-improves. And then we did that with the Twitter poll and the Reddit poll and the Gmail poll and this poll as well. So, instead of having a loop for the entire thing, just like one huge loop, we instead had a bunch of like micro loops nested together inside of one run. That’s graph engineering. So, I did a whole video on that. It can be a little complicated, but at its core, that’s all graph engineering really is. It’s just a bunch of looping agents that also talk to one another. That’s all it is. And
39:55 — for most people, this is total overkill. You don’t usually need this, but conceptually, that’s how it’s working. Now, let’s shift the discussion over to dynamic workflows and ultraode. So, what exactly is ultra code? How does this different from max effort? Well, Ultra Code, what it’s going to do is it’s essentially going to create a custom harness for whatever problem you’re trying to solve. What that means in practical terms is it’s probably going to spin up a bunch of sub aents to deal with whatever issue you have. This can be extremely effective, but this is also going to be extremely expensive. One
40:27 — example of a dynamic workflow is for slashde research. This is essentially like a pre-built dynamic workflow and it is very similar to how deep research works if you’re just on the web app and do deep research. So if I run deep research, what’s going to happen is it’s going to spawn a ton of sub aents and these sub aents are going to do a bunch of different tasks. So if I ask a question like let’s deep research best use cases for dynamic workflows inside of cloud code.
41:00 — It’s going to now instead of doing a standard web search where it spawns, you know, maybe like five sub agents to basically do a Google search, it’s going to spawn several more than that. I’ve had it spawn well over a hundred sub aents. And these are going to do a number of tasks. It’s going to actually go out on the web and scrape data. It’s then going to create adversarial agents that take a look at the data we found and then compare and contrast it to see what actually holds up to scrutiny. Then it’s going to do synthesis and then it’s going to give me a final report. And so you can see over here it decided that for the scope it’s going to decompose a
41:30 — question into five search angles. It says it’s only going to need five parallel web search agents, which is nice for us since we’re on Fable. And then it’s going to pull the top 15 sources, verify everything with a three vote adversarial check on each claim, and then finally synthesize it. And so we can see it working here over on the right. Since we spawned six agents and every single agent has like a pretty much a token cost right off the bat, we’ve already burned 314,000 tokens. So, I wasn’t joking about that. Now, when you run Ultra Code, if you
42:01 — don’t tell it specifically, like let’s say I’m on Fable 5 and running Ultra Code, it’s going to use Fable for these sub agents, which can be a problem because what if it did say, “Hey, I’m going to spawn 100 web search agents.” Now, when I give it the prompt to get away from something like that, you can specifically say, “Hey, limit it to 20 sub agents. Limit it to 50 sub aents.” Or you can say something like, “I want you to use sonnet for the sub aents or opus for the sub aent.” So, you aren’t necessarily a slave to whatever model you’re you’re using at that time. Now, anthropic put out a pretty good blog explaining dynamic workflows. And so,
42:32 — what’s going on under the hood is that it’s writing orchestration scripts that run tend to hundreds of parallel sub aents in a single session, checking its work before anything reaches you. And here’s some examples of different sorts of dynamic workflows. Remember, when you run dynamic workflows in Ultra Code, Claude Code is going to figure out the best one that fits your problem. It might be one of these. It might be something completely different. So, classify and act. You give it some sort of task. We have a classifier agent that then chooses the best sub agent for you. Fan out and synthesize and adversarial review. If we kind of combine these two, that’s what we’re doing with deep
43:02 — research, right? We have some sort of task, find this information. It fans out on the web, gets all the information, and then it also does adversarial review to see what actually makes sense before synthesizing it for us. Then we got stuff like generate and filter. We have a tournament style thing where we try different attempts to solve some sort of problem and includes judges. And then we have loop until done, which again very similar to loop engineering. If we jump back inside here, we can see this deep research run we started earlier included 103 agents and burned six million
43:32 — tokens. And this was all on fable. So, you can imagine how expensive this is. Here’s a look at the actual report it gave us. And as you can see, pretty deep and also includes 21 different sources. And out of all the examples, I think deep research is the one you’re going to use the most. And if you’re someone who’s about to embark on a pretty complicated project and you really want to get all your ducks in a row before you start even going into plan mode, I highly suggest using deep research so Claude Co can go out there and see what’s what before it starts building. Let’s talk a little bit about model routing. And really that’s just how can
44:02 — we bring in outside models into the clawed code fold because some of the big players like chatbt their models are great. Soul 5.6 is awesome. Luna and Terra are extremely token efficient and one sort of thing you need to keep in mind is AI systems in general and these models in general don’t do a great job of grading themselves. So if I ask Claude code to grade its own work, I ask Opus to grade its own work, be able to grade its own work, it’s going to pretty much always say, I did a great job. So how do we solve this problem? Especially
44:33 — when we’re having it grade work that we ourselves can’t grade, it’s like outside our domain of expertise. I don’t really know if that code is great. Well, why don’t we bring in another frontier agent to take a look at our work? And the way we’re going to do this is through different skills and plugins. There’s actually an official Codeex plugin for Claude Code. This is from OpenAI themselves. And this allows you to call on codeex from claude code. You simply take this URL, you paste in cloud code and say I want to install this. You can and from here you can have codeex do adversary review of the code you’ve already created or you can even use this
45:05 — to have codeex work on specific features of your product. And when it comes specifically to the planning stage, I created a skill called grill meex which combines Matt PCO’s groom skill with adversary review from codec. What happens is you and Fable talk to one another. You come up with a plan and then that plan is routed to Codex and Codex and Fable kind of have a back and forth up to five rounds where Codex takes a look at what Fable created. Codex says, “This is wrong. This is why.” Claude reacts to it, says, “Okay, I’ll fix that or I don’t agree.” And
45:35 — they keep going back and forth until they reach an agreement. And so this kind of solves the problem of, you know, these models struggling to grade their own work and gets us a second set of eyes. And so we can feel pretty confident moving forward in complex cases. And this sort of model can be taken even a step further if you’re someone who doesn’t want to use something like Codeex and you instead want to rely on much cheaper models or even local models. The whole point of it is you aren’t stuck just using Opus, Haiku, Sonnet, and Fable. We can bring in whatever we want, especially if we
46:06 — use it as a skill. And again, you would just use the skill creator skill to do that. And last but not least, we have our custom agentic OS structures like you see here and here. These are all about building a custom wrapper on top of Claude and giving us sort of a visual interface that we can’t get elsewhere. And it’s ultimately custom. Now, the real value of these things isn’t the visual sort of wrapper, even though this is, you know, useful for me for me. This gives me my social media metrics. I can easily click and get a deeper dive on that. I get my research done every single day, all in one place, showing me
46:37 — what’s going on in GitHub, hacker news, etc., etc., All these buttons are related to specific skills and automations that I can run on demand. And over here, I pretty much have the exact same thing, but it also has a voice mode. But like I said, the true value isn’t in these cool visual layers. The real value is in the skill architecture behind it all. The whole idea of these things is turning Claude into essentially your personal assistant or an actual worker in your organization that can do everything you do. So for me, what you’re looking at here is
47:07 — essentially all the skills I use in my day-to-day mapped across all the different domains that are important to me. So you know, I have things related to memory, things like productivity, which is like Gmail, my calendar, all that. I have stuff related to research, content, my community, my AI agency, sales, etc., etc. Every single one of these things are a task I would do on my own manually. And instead of doing them on my own manually, I have now mapped them to specific skills. And if they make sense, I turn those skills into automation. And when we look at that in a big picture sense, that is really what
47:38 — an Agentic OS is. It’s a series of skills and automations that you have mapped to your daily tasks. Now, how do you build something like this? Well, we’ve already kind of talked about it, right? This is just building skills. It’s just using the skill creator skill, turning on your microphone, and giving it a stream of consciousness about what it is you do dayto-day, week to week, and then asking Claude code, can we turn those into skills if they make sense? And if they do, you do it. And you do that over and over again and eventually you create this corpus of skills that allow you to automate huge swaths of your life. And behind all that is Obsidian. And Obsidian allows us to
48:10 — easily track everything we do in terms of a markdown file system. Now Obsidian in itself isn’t giving you some like crazy upgrade to what Claude code can do, but allows us to actually track everything. It allows it to give it some memory. And this is sort of the basis of like the Carpathy Obsidian Rag system you’ve probably heard before. Now, as a very quick review of what I mean when I’m talking about an obsidian memory system. This is just a file structure. This is just a coherent file structure that clog code lives in so you understand where things go. This is again often referred to as the Carpathy
48:40 — Obsidian thing. And it’s very simple. You have some sort of folder that your agentic OS lives in. Mine is called the vault. Inside the vault, you have some sort of file structure that looks like this. It doesn’t have to be exact, but the idea is in one of those folders, we have the raw section. This is sort of where raw data goes. This is where research goes. We then have a wiki wiki section and we have an output section. The wiki section is where we take all this raw data and we essentially turn it into different reports or wiki style articles. So imagine I had cloud code go do a bunch of research on AI agents.
49:10 — We’ll dump the raw information here and then it creates an article about AI agents under the AI agents subfolder. Now let’s say I wanted to create something from that AI agent type wiki. I wanted to turn it into a slide deck for example. Well then that slide deck would go to the output section. And the idea is with this simple sort of model, I am able to deal with a ton of files, potentially hundreds of thousands, and it is set up in a clear way that I, as a human being, can easily navigate. And B, Claude can easily navigate. And if it
49:41 — can easily navigate this file structure, it’s going to make it more accurate and ultimately cost less tokens. The key of this whole thing are these sort of like index files at every step of the journey. Every time I go deeper into the file structure, there’s essentially an index. markdown file telling me what is going on inside of there. So essentially like a table of contents. Now I have a ton of content that goes a lot deeper into this if that was confusing. But the idea of these agentic OS systems is that this visual wrapper sits on top of that custom skill architecture I just talked about and is sort of buoied by that
50:11 — Obsidian memory layer which helps you navigate everything you’ve done and helps cloud code navigate it just a little bit better. And the way these systems work because it’s like hey I’m not actually using cloud code here am I? is we use headless claude which uses -p. So essentially a command is sent to the terminal instead of running like just claude it uses claude-p which means claude essentially runs invisibly in the background. There was some drama about this for a while where anthropic was saying they were going to charge different rates for that and wouldn’t be usage but that’s no longer the case. So
50:42 — these sort of systems are actually just as moneyefficient as anything else. Now, creating an agentic or cloud OS like this isn’t 100% necessary, but I think it is necessary to sort of create these sort of skill architectures. That really is the backbone of it all. And of course, if you want to get my exact setup, I have that inside of Chase AI Plus as well. So, that is where I’m going to leave you for today. We covered a ton of different topics, and these are the ones that I felt kind of can give you the most bang for your buck in terms of what you should focus on from
51:12 — everywhere from the beginning all the way to a more advanced level. So, as always, let me know what you thought. Make sure to check check out Chase Plus if you want to get your hands on my Claude Code Masterass, which kind of covers a lot of the same stuff we did here, just in much greater detail. Besides that, I’ll see you