[0:00]For the longest time, I made this huge mistake of thinking that Claude Code was just for coding. And boy was I wrong. Once you start to realize how you can use Claude Code as a second brain, especially with the help of Obsidian and skills, you unlock a whole new world of possibilities, and that's what I want to introduce you to right now. Just giving you a glimpse into how I'm building these systems that literally save me dozens of hours every single week. And it really is just driven by Obsidian, Claude Code and skills, two tools, one capability. Now, if you were to ask me to cover my entire second brain system that we're looking at right here, it would take hours. And so to make good use of our time, I'm just going to quickly cover the capabilities that I've built into my system here. So you can take inspiration from it, and then of course, I'll show you how you can get started building your own second brain. I even have this resource that I'll link to in the description, covering all of the core skills that power my second brain. And fair warning, there is a lot in my system. Take a look at all of these skills and capabilities that I have, and so we really are getting into the belly of the beast. And quite a bit of this is still pretty unrefined, which is why I'm not sharing everything right now, but I certainly will be making more content as I build this all out. And this next part, it's a little corny, so bear with me on this. But having this system, legitimately makes me more excited to wake up every single day because it helps me ideate and keep my ideas organized. And so I just feel like things run so much smoother, and I want that for you as well. And that really is the point of a second brain. It's not to automate our business, replace us or generate a bunch of AI slop to post all over social media. That is not what I'm doing here. Instead, the main point is to help me ideate, stay organized and research things for me. And so all of the capabilities that we have here, and the ones that I'll share in the template, they fall under one of those three categories. And so even for like the X-post skill, or the YouTube script or LinkedIn post, it's not like it's generating something for me to just post right away. There's still a lot of human touch, but it's augmenting me, helping me every single day I wake up, with taking all of my ideas, supplementing it with research, and then finding ways to help me frame it, for example.
[3:16]So with all of this, Claude Code, it helps me do a bunch of research on the latest AI news, even giving me content ideas based on what I've found from all the sources that I have curated for it. It helps me generate the Excalidraw diagrams you see on my channel all the time. YouTube scripts, this is very important for me. There's the new Remotion skill that allows Claude Code to generate videos, things like B-roll that I can include for videos on my channel or posts on other platforms, really, really powerful. I also have a skill that allows me to connect MCP servers like Zapier without taking up tens of thousands of tokens up front for the context of Claude Code. So I'll show you this later as well, all the other different kinds of capabilities so I can pull things in my email or Asana and link it to documents and research that I have in Obsidian. And one of my favorite skills that I'll show you in just a bit here, this is one of them included in the template is this one to generate on brand PowerPoint slides. This matches all my branding for Dynamus, and I'll show you how you can brand it to yourself as well. This is a lot more powerful than other AI PowerPoint tools like Gamma or the skill that Anthropic released for us.
[5:12]So, there are two things that I want to cover with you right now. First of all, I want to explain more why Claude Code, Obsidian, and skills is such a powerful combination. And then with that, we'll get more into the template because I want to show you how in just like 5 to 10 minutes, you can get some of these skills up and running for yourself. And I promise you that if you try this, even just give it a little bit of time, you'll be very inspired to start building a second brain for yourself. And I'm going to be very concise on purpose in this video. The point is not to dive really deep into what I've built for myself, but just get your brain juices flowing for how you can build your own second brain.
[6:12]And when you start to use these skills, you really will start to see dozens of ideas for how you can use these capabilities to help you ideate, research and organize. And if you want to dive even deeper into everything that I cover in this video, definitely check out Dynamus. In my community, I did a workshop just this Friday where I go even deeper into my second brain, even the entire research engine that I have powering a lot behind the scenes. Plus in Dynamus, I have the new agentic coding course. So a lot of value here, I'll have a link in the description. All right, so now let's talk about the beauty of this combination. Now, the first big question you might have is, Cole, if code is literally in the name Claude Code, why are you so obsessed right now for using it for everything but coding? Well, my friend, I'm glad that you asked. Let's talk about that. So, first of all, everything that I'm going to cover here very briefly, applies to any coding agent. It's just that Claude Code is generally considered the best at the agentic loop with reasoning in general.
[7:31]And so when we think about what a coding agent needs, the capabilities that we're giving it, it really comes down to these five categories. It needs to work with our files, search through them, run terminal commands, search the web, and then it needs access to code intelligence. And so this one right here is the only one that really makes it specifically a coding agent. Because if we take this out of the picture, then all we have is just a very powerful general assistant that is running on our machine. And that's exactly what we need for a second brain. So let's think about our big three again. For our second brain, we want to use it for research, ideation and organization. And so for research, we have web search and terminal execution. And then for our ideation, generating ideas, and for our organization, we have terminal execution, file operations, and searching through our files. Especially with Obsidian, like I'll talk about in just a second here, this is all that we need. Especially because Claude Code is built to be able to handle these longer running, more complex tasks. Coding is not the only complicated thing. For other use cases in our life, we need to take a larger task, break it up and knock it out one at a time, just like we have with a lot of our coding tasks. So when we think about research or just working through a larger idea and thinking about different ways to frame it, whatever it might be, all the skills that we're covering here in this video, Claude Code is just so good at all of it. And all of this is even more powerful when we pair it with Obsidian, because Obsidian is our local and private knowledge base. All of our files are simply stored as markdown docs on our computer. So think about this, markdown is one of the best formats for LLM understanding, and Claude Code does a really good job with our file system. And Obsidian is both, it's really good format, it's in our file system, so Claude Code does such a good job helping us organize our ideas, which we can see in the graph in Obsidian, and helping us search through everything. And so, really Obsidian is my canvas. Everything that my second brain generates, like PowerPoints and YouTube scripts and ideas for post for different platforms, I manage it all here. This is where I add my human touch and really work on all the content because like I said, in the end, my second brain is just helping me with the ideation organization. I'm the one who's really producing everything in the end, and I do it in Obsidian. And it's also much better than something like Notion because then we have to use an MCP server. We can't just have Claude Code work directly in our file system, like we can with Obsidian. Now, the last part of our combination is skills. Why are agent skills so important for our second brain? It's really what brings everything together because this is how we give all of our own capabilities, processes, and guidelines to Claude Code. And we're doing it in a very flexible and context efficient way. And the context efficiency is really the most important part because if we did something like MCP servers instead, then as our second brain is growing and we're adding dozens, maybe even hundreds of capabilities someday, then we're completely overwhelming the LLM with context. Because with MCP servers or any other way to give tools to our agent, we have to describe all of the tools up front. But skills implement progressive disclosure. I'll talk about this in a little bit, and so it knows about all these capabilities, but it only loads them in when it actually needs to leverage them based on something that we have requested. And then also makes it very flexible because Claude Code can essentially specialize itself in each session depending on what I wanted to do. Like if I wanted to help me ideate for a LinkedIn post, then it's going to load in the skill, and those are going to be its primary instructions. Or if it's helping me with my next YouTube video, maybe it's going to load in the content engine and the YouTube script skill, and then it'll be specialized in that way. Now, sometimes you do want to use MCP servers still. So I even have you covered for that. If you want to incorporate MCP servers into your second brain to work with your different services, then I have a skill that basically instantly transforms any MCP into a skill. And as a part of the Readme here, I have instructions for how you can set that up. Very easy to get started, because like, for example, I have the Zapier MCP server to connect my second brain to my Gmail, Google Calendar, Slack, and Asana. And by the way, with Zapier, you also can select the individual tools that it has, and so it's only read operations. I'm not trusting my second brain to send emails on my behalf, for example. And so I can take the Zapier MCP and I can use this skill, one of them that I have in the template for you, so that my second brain has access to it. But it's not like it has to load in the context for all 20 sum of these tools that we have right here. There's a lot of information if I was just connecting this MCP directly. So we're sticking with skills for everything, even bringing in MCP. So, MCP implements the idea of progressive disclosure. Basically, instead of dumping all of the capabilities into our agents context window up front, we are allowing it to discover things based on just a short description that we give it right away. And so, for example, with the PowerPoint presentation skill that I'll demo for you in a little bit here, this is one of them that's in the template. We just have this short description that essentially tells Claude Code that you want to leverage this skill whenever the user is asking to create or edit PowerPoint presentations. And so the description is the only thing that's given to Claude Code up front. But then it knows based on that, like, okay, the user just told me that they want to generate a PowerPoint, and I have this description in my prompt. Now, I should probably go to the next level of progressive disclosure, and so for skills, that is loading the skill.md file. So for any skill, this is the primary file that gives the main instructions for this capability. And so it'll load in this entire skill.md. So where the description is at the top, this is the file that it'll now read the entire thing when it wants to leverage the skill. And then within the skill itself, we even have other files. This is the third layer of progressive disclosure, because sometimes even when you're leveraging the skill as a whole, there is even more specific information that you want to call out so the agent can reference it if it needs. But you don't always need that when you're using the skill. So, for example, with the PowerPoint generator, we have our primary skill that MD that I was showing you earlier. But then we also have this cookbook folder, so a lot of different Python scripts. We're actually using code execution to edit and create our PowerPoint slides, which makes it very powerful and flexible. I'd highly recommend trying out this skill in particular, I think it'll blow your mind. And so sometimes we want to use the Python scripts for generating slides, sometimes we want to use the Python scripts for editing slides. It depends on the request, what we're doing in the current session, that's why we want to split these out into other either reference markdown documents or Python scripts that we can use when we're leveraging this skill. And so in the end, it's just very, very flexible, giving only the context that Claude Code actually needs. And that's what makes it possible for us to give so many capabilities to our second brain.
[18:08]All right, so I hope at this point you already have a ton of ideas to build your own second brain. We've covered what's possible, I've showed you why this combination of Claude Code, Obsidian and skills is so powerful. Now, I want to show you some of these skills that I have for you out of the box. So again, link to this repo in the description, you can take this URL and give it to Claude, ask it to clone for you, and you can use other coding assistance if you want as well. Or you can just go ahead and clone it yourself. So then, you'll have the repo running locally. And so this has a .claude/skills folder. You can take it and use it right here. You can bring it into anywhere else that you're running Claude, and it'll instantly have access to all of these capabilities and I'll explain every single one of them in the Readme here. So right now, I'm not going to cover every one of these skills with you. You can explore them yourself, I don't think that's worth our time right now. I want to just cover the most important ones to get you started, especially the brand and voice generator because this skill is going to guide you through creating the context that makes the second brain really specific to you. So you can use your brand and voice that you generate here, as well as the skill creator, this actually comes straight from Anthropic, to build all your skills going forward and make them really customized to you. Like, this is how I want to create Excalidraw diagrams, or how I want to create PowerPoints. And so I also want to cover the MCP client skill, because there's probably a lot of MCPs that you want to bring over into our second brain. And then we have the PowerPoint generator, which I'm very excited to show you. So, these are the ones we'll hit on right now, starting with the brand and voice generator. So, in order to use any of these skills, all we have to do is load up Claude Code when we have the skills directory with all of these in our .Claude folder. So, I'll just open up a terminal right here, just Control J in Visual Studio, and then I have my Claude already running. And so, starting with the brand and voice generator, we just have to give one of these triggers. These are just examples, pretty much anything saying like, help me create my brand system, it's going to load in the skill. Because remember, we have the skill.md with progressive disclosure here where this description is given to Claude immediately, as long as we have this skill in the our .Claude/skills folder where we start Claude Code. And so if I say, help me create my brand system, I send this in, and then immediately here we'll see that it loads the brand voice generator skill. Take a look at this. There we go. All right, loads the skill, and then that brings in all the instruction which includes a workflow. So, it's going to guide us through a bunch of questions to then generate our own brand. And even to help you as a resource here, I have my Dynamus brand. You can see like what my tone of voice is, what my brand system is. I use this as a really core part of all of my skills that I've generated going forward. And so back over to Claude Code, you can see here that it kicked off the workflow that I lay out in the skill.md for the brand generation. So, I'm not going to go through this full process right now, it's really, really straightforward. You just answer all of its questions, and then within the brands folder right here in the PowerPoint skill, it'll create a brand for you, just like I have mine for Dynamus. So the PowerPoint skill in particular really leverages this to make sure the slides are personalized to you. That's why I have it here. You could also change it if you want to make it so there's some separate folder where it stores all of your brand information. So, it's entirely up to you. You really can customize all these skills to whatever you want, and I really encourage you to do that to make it super specific to you. You can also read through for every single one of the skills here, I've got more information in these chevrons that you can expand. And then I have the triggers listed out as well. And so now with our brand and voice created, we can make more skills, or we can go right into one of the ones that I have in the template for you. So we can start with the PowerPoint generator. So, I have this template, all the skills here I have in my personal engine, so I've got the PowerPoint generator here as well. And it is literally as simple as this. I'm going to go into my speech to text tool, watch this.
[24:08]I want you to generate a presentation on why Claude Code, Obsidian, and skills is the best combination ever. Let's just do like 10 slides in total. I want you to search the web so that you understand all these tools and why they work so well together. Boom. All right, I'll send this in. That is all I have to do. I don't have to tell it to use the PowerPoint skill. I don't have to invoke any kind of command. It understands when it should load things in dynamically. It is a beautiful thing. And so I'll actually show you the presentation after I go through the full process here. And take a look at the presentation that it generated here. This is a fantastic starting point, and I still want to apply my human touch and refine the content and the text and everything. There's probably some room for more variety here at the slides as well. But overall, this saves me so much time. And everything is on brand for Dynamus. It'll be the same for you if you do the brand and voice skill and then you create your PowerPoints after. So the last capability I want to show you really quickly is the MCP2 skill, and this one is really powerful because there are a lot of tools we want to connect to Claude Code, but there isn't really an easy way to do it besides through MCP. And so we can directly turn them into skills so we're not bloating the context for Claude Code and still being able to give it a lot of different access to different tools. And so for this skill.md, essentially, it's just instructions for how it can use this Python script to interact deterministically with different MCP servers. So instead of connecting MCPs directly to Claude, we're interacting with the servers through our Python script. And so it can invoke this with different parameters to list out our servers, the tools in a specific server, call a certain one with the arguments that it supplies. It's very, very flexible how it interacts with any MCP server that we want. And so all you have to do, and I explain this in the Readme as well, is take this example MCP config, make a copy and rename it MCP-config.json. And then you can connect any server that you want here. I just have a couple of examples, but any protocol is going to work, and this is just the exact same JSON configuration as something like Claude desktop. So it should be very familiar to you if you've used MCP before at all. And then also in the Readme, I go through a couple of other details that are pretty important. Generally, MCP servers because they expose so many different tools and have all these different parameters, they can be kind of confusing to coding agents, even the best like Claude Code. And so, generally, it's helpful to, when you first set up your MCP servers, you ask Claude specifically to go through all the tools and document any gotchas in the global rules, like for Claude Code, our claude.md. So, I actually did that myself. So, over to my personal engine here, my personal second brain. I have my global rules where I speak to the MCPs that I have, and any kind of gotchas that I noticed as I was using them. And so you can sort of improve your system over time, so it's more reliable leveraging your MCPs. And so I love using Zapier to connect really quickly to all my different services, like Gmail, Asana, Google Calendar, and Slack. And so this is a pretty integral part to my second brain, being able to reach out to these other services. Because maybe I want to pull something from Gmail and then ideate about that and documented in my Obsidian for example. So I can do that using the Zapier MCP.
[29:13]So, for example, back in my Claude Code, I can use my speech to text and say, I want you to look in my Google Calendar to see what meetings I have on Monday, January 26th, and also what tasks I have as a to-do in my to-do project in Asana. Send this in. I don't even have to say Zapier specifically because it knows from my global rules in the skill that's what it needs to do. So, check this out. Boom, take a look at this. I got my full calendar laid out here, all my meetings and when they are, and my Asana tasks that I got due on Monday, and it even told me what I have overdue, because I got to get this video out right now. And then what I'm going to be doing tomorrow as well. Super cool. We're just scratching the surface here for what is possible with MCPs in our second brain and just with the capabilities in general. So, that, my friend, is everything that I got to cover with you right now on how you can create your own second brain. I hope that this combination of Claude Code, Obsidian, and skills makes a lot of sense. And please leverage this as a resource, a launching pad for you to bring in some skills and start building your own brand and voice and capabilities for your Claude Code. If you appreciate this video and you're looking forward to more things on building a second brain and agentic coding, I would really appreciate a like and a subscribe. And with that, I will see you in the next video.
