--- title: '7 Secrets to Using Claude Code 10x Better' source: 'https://youtube.com/watch?v=wWHL5FSDoAI' video_id: 'wWHL5FSDoAI' date: 2026-07-24 duration_sec: 709 channel: 'Negócios em Mente' --- # 7 Secrets to Using Claude Code 10x Better > Source: [7 Secrets to Using Claude Code 10x Better](https://youtube.com/watch?v=wWHL5FSDoAI) ## Summary This video reveals seven secret commands and techniques to enhance productivity when using Claude Code. It covers commands like Insights, Clear, Ultra Think, BTW, Loop, Schedule, and the open-source skill Caveman, each designed to optimize usage, save tokens, and improve response quality. ### Key Points - **Insights Command** [00:31] — Typing '/insights' analyzes your last 30 days of conversations, providing a customized report on productivity, negative patterns, and areas for improvement. It acts like a consultant giving blunt feedback. - **Clear Command** [03:07] — Typing '/clear' deletes the entire session history. This is useful when switching tasks to avoid context overload, which degrades response quality. It also saves tokens. - **Ultra Think** [04:28] — By including 'Ultra Think' in your prompt, you force Claude to use maximum reasoning effort regardless of default settings. Use it for complex tasks, but be aware it consumes extra tokens. - **BTW Command** [05:48] — Typing '/BTW' followed by a question allows you to ask a quick question without interrupting the main task. It's useful in long sessions where you don't want to lose context. - **Loop Command** [06:42] — The '/loop' command schedules tasks to repeat every minute to every hour, for up to three days. It runs on your computer, so it stops if you close the session or turn off your device. - **Schedule Command** [07:37] — The '/schedule' command runs tasks in the cloud, so it continues even with your computer off. Minimum interval is 1 hour, with daily limits varying by plan (Pro: 5, higher plans: 15-25). - **Caveman Skill** [09:25] — Caveman is an open-source skill that reduces output tokens, making responses more concise. It can be installed by typing 'I want to install the open-source skill cave' in Claude Code. ### Conclusion Mastering these seven commands can significantly improve your efficiency with Claude Code, helping you save tokens, maintain response quality, and automate tasks. Run Insights monthly and use Clear, Ultra Think, BTW, Loop, Schedule, and Caveman strategically. ## Transcript secret commands will make your life much easier. It doesn't matter if you've just started using Cloud Code or if you've been using it for a while. I'm sure you're not familiar with at least one of these commands, and they'll greatly improve your experience. My name is Tainara Shimite and welcome to the Business in Mind channel. Here you'll learn everything you need to know about artificial intelligence. Let's start with one that few people know about, which is the insights command. All you need to do to run it is simply type slash inside. When you type this, Cloud analyzes all your conversations from the last 30 days. It will literally analyze how well you are using the cloud, how productive you are, what is holding you back, negative patterns, and things that need improvement. And that's the most important detail. He will tell you what is working very well in your sessions, what is holding you back and how it could be improved, as well as features and ideas that you can implement to be more productive. Think of it this way: it's like a specialist consultant looking at everything you've done in the last month and bluntly telling you, "Look, you're wasting potential here, here, and here." So, instead of receiving a generic guide, things you could do with the cloud." When you run this command on websites, the cloud analyzes and generates a specific report for you, tailored to your usage. So, it's customized. And these ideas are based on tasks that you are actually trying to do or have already done. So, he won't give you irrelevant and generic ideas. Look, I recommend running this command at least once a month, just to give you a reality check on how well you 're using the Cloud or how poorly. And since we're talking about using the Cloud for what it's really good at, when it comes to creating a beautiful visual, a presentation, or a carousel for Instagram, it's simply not the right tool. And that's where viso comes in, the visual arm of your AI workflow. It transforms text into design ready for any channel. It's perfect for content creators, marketing professionals, or anyone who simply wants to stop wasting time on design. Copy, PDF, any text you've already created. Viso transforms the design into the format you need. Presentation, infographic, newsletter, all in one click. Unlike Canva, Visio handles long documents very well, and you don't need to start building your design from scratch. The design is ready, and you can edit whatever you want if you want. And for those who use the nano banana, the benefits go even further. You have complete control over text, layers, elements, and everything you create is already licensed for commercial use. Are you curious? Visa has a free plan and doesn't require a card. The link is in the description and in the pinned comment. seems simple, but there's a logic behind it. He will delete the entire history of the current session. And the most obvious question is: Tai, why would I want to delete an entire conversation that I just built? Because recent research shows that the more accumulated context, the lower the quality of the response that IAT delivers. It's like trying to think in a completely messy room. Your mind gets overwhelmed with things that have nothing to do with what you need to think about right now. A practical example. Imagine you spent the entire morning using the Cloud to analyze a sales spreadsheet with all the numbers, data, and context of that business. And then, in the afternoon, you decide to create some marketing materials for the company, such as a video script. If you stay in the same session, the Cloud will load all that context from the sales spreadsheet, and that ends up being disruptive. So you're going to run the command with a clear slash, and then you'll start from scratch, and the response will be more accurate and cleaner. And a very important detail is that this also helps to save tokens, which is basically the currency used in the cloud. So, if you're on a plan with a limit, this command helps to stretch your balance. Go ahead and like this video, subscribe to the channel to stay updated, and leave a comment below telling me if you already use Cloud Code and which of these commands you were unfamiliar with. Now let's talk about Ultra Think. This one doesn't start with a slash; you'll just type the text into your prompt. What it does is force the cloud to use maximum reasoning, regardless of its default configuration. It's like you're telling it: "Look, stop everything you're doing and think very, very calmly, because this is a very important task." So, even if your default setting is low reasoning effort, it will use maximum effort to perform that task. This way, you can do all your work in standard mode, maybe even medium, but when you reach a very large prompt, a very complex task, you can simply use this command. It will make that change in reasoning effort only for that prompt. And this is also very useful for saving tokens, especially if you're on the 20-per-month plan. So, only use it when you need the cloud to think deeply about a more complex or technical task. Remember that this consumes extra tokens, so only use it when it's really necessary. Now let's move on to the next one, which is one of my favorites, which is... By the way, but we'll just write BTW. but we'll just write BTW. So, slash BTW. You've probably been through this. The cloud is there working on a very large task, writing, processing, generating... Something extensive, creating something complex. And then suddenly a quick question pops into your head, but you don't want to open a new session because you want to use the context of that conversation, but what's happening there. And this command solves the problem. You will type /BTW followed by your question and the Cloud will answer your question while continuing to work on the main task. It's the equivalent of sending a quick message to a colleague while they're in the middle of a meeting, and then they continue in the meeting, but they pick up their very useful, especially when you already have a good base of context built in that session and you don't want to lose that just to clear up a small doubt. Now let's get into real automation and talk about the loop. It allows you to schedule tasks to repeat automatically, every minute and even every hour, but for a maximum of three consecutive days. For example, let's suppose you want to be notified whenever important news related to your work area comes out. You can configure a loop. Every 30 minutes, the Cloud searches for information on specific websites and delivers it to you. Summary: It will run automatically while you work on other things. But there's an important detail. The loop runs on your computer. This means that if you close the session or turn off your laptop or computer, it will stop. Also, after three days it automatically stops, but for most day-to-day automations, that's more than enough. Okay, Tai? But what if I need an automation that continues long-term and not just for three days? And that's exactly where our next command comes in: `schedule`. It does something very similar to what `loop` does, but with important differences. First, it runs in the cloud and not on your computer. This means that even with your computer turned off, it will continue executing tasks at the scheduled time. The downside is that it doesn't access local files on your computer, unlike `loop` which can read your files normally. In `loop` you can configure intervals starting from 1 minute. Here in `schedule` it's from 1 hour, that is, at least every hour. It's ideal for automatic daily reports, recurring team tasks, anything that... This needs to happen regularly, without you needing to remember or be online. Just be aware of the daily execution limits . In the Pro plan, it's five per day, in the next plan it's 15, and in Team Enterprise it's 25. These numbers, of course, Depending on the day you're watching this video, it may have already changed, but it's important to keep in mind that there is a limit. Now, if you want to get ahead and know the latest news from A before everyone else, then I invite you to join our free WhatsApp group. And don't worry, there are no good morning messages or stickers there, just news manually selected by our team. Just point your phone at the QR code that's appearing on the screen or click on the pinned comment. There you'll also find our free materials, such as courses, prompts, and spreadsheets, as well as exclusive discounts and bonuses on tools we use and recommend. So, if I were you, I wouldn't miss this opportunity. Go ahead and take a look. Right now, I want to share Caveman Man with you. This This isn't a native Cloud command; it's an open-source skill , meaning anyone can install it for free. To install it, simply open Cloud Code and type: "I want to install the open- source skill cave." It will install it automatically. Then, you just use barracks in your commands. What it does is significantly reduce the number of output tokens, which are the tokens used in Cloud responses. A simple analogy: tokens are like fuel. Every time Cloud responds, it uses a certain amount of fuel. Cave makes Cloud more concise and direct in its responses, using less fuel without losing important content. Again, if you 're on a plan with a usage limit, this skill also greatly helps to stretch your usage throughout the month. Now, a summary to help you solidify everything you've learned: Insights. Analyzes your actual usage over the last 30 days and delivers a report on how you can be more productive. Run it at least once a month. Clear. Clears the session context. Use Ultra Think when switching tasks to maintain response quality and save tokens. Force the cloud to use maximum reasoning on a specific prompt. Use it for complex tasks that require deeper reasoning, and remember that it consumes more tokens. By the way, or BTW, allows you to ask a quick question without interrupting the task the cloud is already working on. It will be your greatest ally in long sessions. Loop schedules recurring tasks on your computer with a minimum interval of 1 minute and lasts only 3 days. Schedule: Cloud automation, doesn't depend on your computer being on. Minimum interval of 1 hour and usage limit varying according to the plan. Cavem: Open-source skill that reduces token consumption in responses. Ideal, essential for those on plans with usage limits. However, I must confess that my favorite cloud feature is coworking. And in this video, I teach you how to use it in practice, from how to use it in practice, from beginner to advanced levels.