I will include my information on how to get the most out of Claude Code. Claude Code is a powerful coding-line interface (CLI) tool. You can open it directly in your terminal and start coding.
This makes it a useful tool for both experienced and end-to-end coders. However, it also works well for non-technical people, as they can write by typing in natural language, and they don't have to interact with a complex IDE.
In this article, I will cover why you should consider using Claude Code to enter codes, highlighting its effectiveness in running code without manual reviews. In addition, I will cover some techniques I use to get the most out of Claude Code, before highlighting some limitations of Claude Code.
I'm not affiliated with Claude Code, and I'm just writing this article from my experience using the tool.
Why use Claude Code
If you want to practice coding these days, you should embrace agent coding. I think this applies whether you are new to programming or if you have been programming for 10+ years. We all have to be agents who use code to be successful developers.
Now you have a huge variety of such tools to choose from. Codex Claude is one of several CLI tools, and other options are OpenAI Codex or Gemini CLI. These CLI tools are great when you don't need to look at the code you're building, usually for simple implementations or debugging.
However, I am very aware that my coding agents are able to use a large scale with fewer and fewer errors, which is remarkable. I believe that we will soon be in an era where coding agents do all the coding (you could also argue that we already are), and in such a scenario, CLI tools will be powerful, as you never need to look at any code.
If you want to look more into code, however, you might consider an integrated coding experience, in tools like Cursor or Antigravity, which are forks of VS Code, but with agent functionality deeply integrated into the application.
I like to use both CLIs and IDEs for agent coding. When I want to code without doing manual code reviews, I tend to use Code Claude more and more. I think the situations where you should strongly consider using the Claude Code are:
- You are correcting a mistake
- You search the production logs
- You use a feature with easy or medium difficulty
If you are in any of these situations, or similar, I strongly urge you to try Claude's Code, or any other CLI code tool, and see how well it works.
Techniques I use
In this section, I'll cover some of the techniques I use to get the most out of Claude Code. Each subsection will cover a specific method, including why I think it is important.
Slash Commands
Slash commands are commands that you save for easy access. You usually do this in a command that you run regularly, such as:
- Investigates production logs
- Creating PR
- Checks if the code is ready for production
When I realized I was using the second prompt, I created a slash command at the prompt. This makes it much more efficient to run the prompt in the future, as I can just call the command starting with a slash (/), instead of typing the prompt every time.
The main reason for using slash commands is to save time and consistency. Writing content takes time, of course, and besides, you often won't be consistent when you write the same content multiple times.
For example, if you have information to check that the code is ready for production, you might have a checklist for the agent to go through, such as:
- Searches for missing species
- Running a pre-commit test
- It scans the Docker image
However, if you type information all the time, you run the risk of forgetting any of these points, making the agent ineffective.
User memory and project memory
I am also a heavy user of both user memory and project memory in Claude Code. User memory is memory that applies to all instances of your Claude code, while project memory is memory that only applies to the folder and subfolders you're currently working on.
I use user memory whenever I want Claude to remember some preferences I have that apply to all collections. This is an example:
- How to create good PR
- How to view production logs
- Information about my IaC stack
If I have preferences that only apply to a specific project (or at least that don't apply to all my coding projects), I store them in the project's memory, in a file called AGENTS.MD. I usually use AGENTS.md instead of CLAUDE.md, because there all my coding agents will find the file, for example, if I decide to try Cursor or other tools in the future.
Using memory is important, as it prevents Claude Code from making the same mistake twice. If Claude used the wrong table name or checked the wrong log group, you can store in his memory what he did wrong, along with the correct table name and log group name. If you interact with Claude Code in the future, it will remember this information and work properly next time.
If you keep doing this for every Claude Code error, it will be personalized to your codebase and your preferences and will work much better for you as a coding agent.
Editing mode
Edit mode is also incredibly important when using Claude Code. Editing mode is a mystery killer. If you tell Claude Code in program mode, it will analyze your information and codebase and ask questions about anything that is unclear or ambiguous.
Editing mode is a mystery killer
This is very useful, as it is close to impossible to write clear information, and having Claude's code identify the unclear parts of your information and ask you questions so it is a good way to align how the coding should be done.
I probably start 90% of my Code Claude sessions in program mode; the only exception is when I'm doing a very simple fix or just writing test code. I recommend that you try the editing mode as much as possible, and only use it when you feel it is not necessary.
Do everything with Claude
My final tip is that only your imagination limits what you can do with Claude Code. For example:
- Do you want to make a presentation? Ask Claude to give a presentation on Latex Code
- Want to analyze CloudWatch logs? Give Claude AWS permissions and have it analyze it for you
- Do you want to create a portfolio website? Have Claude block it and send it to Vercel. You don't even need to leave the terminal
The possibilities are endless, and I strongly urge you to use the following mindset:
Whenever you encounter a problem, you should think: How can I solve and automate this using Claude Code
This attitude will save you a lot of time, as you will see that many tasks can be solved by giving Claude the necessary permissions, and let him try to solve the problem for you.
Challenges with Claude Code
I have mentioned many good sides of Claude Code in this article, because it is a really good tool that makes me more successful as a developer. However, there is also something wrong with the Claude Code.
The first problem I have with Claude Code is that it is slow at times, especially when compared to using the Cursor. This happens even though I use Claude Sonnet 4.5 or Claude Opus 4.5 in Cursor.
I'm not entirely sure why this is the case, but I often find that when I'm analyzing input groups or making simple adjustments, Claude in Cursor is simply faster than using Claude Code.
Another disadvantage of Claude Code is that it is based on the end. If you want to understand your code better and review it, I would encourage you to use an IDE instead. It should also be noted that there is a Claude Code extension for VS Code, which can be installed in any fork of VS Code, such as Cursor.
The conclusion
In this article, I talked about Claude Code, covering why and when you should use the tool. I highlighted that Claude Code is the best for coding where you don't need to check the code manually, which is often the case with rapid development of LLMs. I then put together some techniques I use to get the most out of Claude, such as using more memory and edit mode, and trying to use Claude's Code for almost every problem I encounter. I believe that coding agents are here to stay, and that you should embrace the technology as much as you can. You should always test the limits of what these coding agent tools can do, in order to be the most efficient developer possible.
👉 My Free Resources
🚀 10x Your Engineering with LLMs (Free 3-Day Email Course)
📚 Get my free ebook Vision Language Models
💻 My webinar on Vision Language Models
👉 Find me on social media:
📩 Subscribe to my newsletter
🧑💻 Contact
🐦 X / Twitter
✍️ In the middle
