, I’ll cowl my expertise on how one can get probably the most out of Claude Code. Claude Code is a strong coding command-line interface (CLI) instrument. You possibly can open it straight in your terminal and begin coding.
This makes it an efficient instrument each for proficient coders acquainted with a terminal. Nevertheless, it additionally works properly for non-technical folks, since they’ll code just by typing pure language, and don’t need to work together with an advanced IDE.
On this article, I’ll cowl why you need to think about using Claude Code for coding, highlighting its effectiveness at implementing code with little to no guide evaluation. Moreover, I’ll cowl some particular strategies I make the most of to get probably the most out of Claude Code, earlier than highlighting some limitations of Claude Code.
I’m not affiliated with Claude Code, and I’m merely writing this text from my expertise utilizing the instrument.
Why use Claude Code
If you wish to follow coding these days, you need to embrace agentic coding. I feel this is applicable whether or not you simply began programming or in the event you’ve been programming for 10+ years. All of us must be leveraging coding brokers to grow to be extra environment friendly engineers.
Now you could have a big number of such instruments to select from. Claude Code is considered one of a number of CLI instruments, with another choices being OpenAI Codex or Gemini CLI. These CLI instruments are good in the event you don’t want to have a look at the code you’re creating, usually for less complicated implementations or bug fixes.
Nevertheless, I’m noticing increasingly that my coding brokers are capable of implement bigger scale implementation with fewer and fewer errors, which is noteworthy. I imagine we’ll quickly be in an period the place coding brokers are doing all coding (you can additionally argue we’re already there), and in such a situation, CLI instruments shall be highly effective, because you by no means have the necessity to have a look at any code.
Should you do need to look extra on the code, nevertheless, you’ll be able to think about extra built-in agentic coding experiences, in instruments akin to Cursor or Antigravity, that are forks of VS Code, however with agentic performance deeply built-in into the appliance.
I like utilizing each CLI’s and IDEs for agentic coding. If I need to code with out performing guide opinions of the code, I have a tendency to make use of Claude Code increasingly. I feel eventualities the place you need to strongly think about using Claude Code are:
- You’re fixing a bug
- You’re investigating manufacturing logs
- You’re implementing a straightforward or medium problem characteristic
Should you’re in any of those conditions, or comparable, I strongly urge you to check out Claude Code, or every other CLI coding instrument, and see how properly it performs.
Methods I exploit
On this part, I’ll cowl among the strategies I make the most of to get probably the most out of Claude Code. Every subsection will cowl a particular approach, together with why I feel it’s vital.
Slash instructions
Slash instructions are prompts that you just retailer for straightforward entry. You usually do that on a immediate you run regularly, akin to:
- Investigating manufacturing logs
- Making a PR
- Checking if code is production-ready
The second I understand I’m operating a immediate for the second time, I create a slash command with the immediate. This makes it way more environment friendly to run the immediate sooner or later, since I can merely name the command beginning with a slash (/), as an alternative of typing out the immediate each time.
The principle cause for utilizing slash instructions is time financial savings and consistency. Writing out prompts takes time, after all, and moreover, you’ll typically be inconsistent when writing out the identical immediate a number of occasions.
For instance, if in case you have a immediate for checking if code is production-ready, you might need a guidelines the agent ought to undergo, akin to:
- On the lookout for lacking sorts
- Working pre-commit checks
- Spinning up the Docker picture
Nevertheless, in the event you’re typing out the immediate each time, you danger forgetting any of those factors, which then makes the agent much less efficient.
Person reminiscence and mission reminiscence
I’m additionally a heavy person of each the person reminiscence and mission reminiscence in Claude Code. Person reminiscence is the reminiscence that applies to your whole Claude code classes, whereas mission reminiscence is the reminiscence that solely applies to the folder and subfolders you’re at the moment working in.
I make the most of the person reminiscence every time I would like Claude to recollect particular preferences I’ve that apply throughout repositories. That is for instance:
- Methods to create good PR’s
- Methods to verify manufacturing logs
- Details about my IaC stack
If I’ve preferences that solely apply inside a particular mission (or at the very least that don’t apply to all my coding tasks), I retailer them within the mission reminiscence, in a file referred to as AGENTS.MD. I have a tendency to make use of AGENTS.md as an alternative of CLAUDE.md, as a result of then all my coding brokers will discover the file, for instance, if I resolve to attempt Cursor or different instruments sooner or later.
Using reminiscence is vital, because it prevents Claude Code from making the identical mistake twice. If Claude used the unsuitable desk identify or checked the unsuitable log group, you’ll be able to retailer in its reminiscence which mistake it made, and the proper desk identify and log group identify. Whenever you work together with Claude Code sooner or later, it would then keep in mind this data and be extra environment friendly subsequent time.
Should you proceed doing this for all of the errors Claude Code makes, will probably be personalised to your codebase and preferences and shall be far simpler for you as a coding agent.
Plan mode
Plan mode can also be extremely vital when utilizing Claude Code. Plan mode is basically the paradox killer. Whenever you immediate Claude Code in plan mode, it would analyze your immediate and codebase and ask questions on something that’s unclear or ambiguous.
Plan mode is the paradox killer
That is tremendous useful, because it’s near unimaginable to write down a totally unambiguous immediate, and having Claude Code determine unclear components of your immediate and ask you questions on it’s a good technique to align on how a coding implementation must be finished.
I in all probability begin 90% of my Claude Code classes with plan mode; the one exceptions are after I’m doing a little tremendous easy fixes or just writing experimentation code. I like to recommend attempting plan mode as a lot as attainable, and solely not utilizing it whenever you really really feel it’s not vital.
Do every part with Claude
My final tip is that solely your creativeness limits what you are able to do with Claude Code. For instance:
- Do you need to make a presentation? Ask Claude to make a presentation in Latex Code
- Do you need to analyze CloudWatch logs? Give Claude AWS permissions and have it analyze for you
- Do you need to construct a portfolio web site? Have Claude vibe it up and deploy it to Vercel. You don’t even have to depart the terminal
The probabilities are countless, and I strongly urge you to make use of the next mindset:
Everytime you encounter an issue, you need to assume: How can I remedy and automate this utilizing Claude Code
This mindset will prevent a number of time, because you’ll understand that a number of duties may be solved by merely offering Claude with the required permissions, and let it try to unravel the issue for you.
Challenges with Claude Code
I’ve talked about a number of optimistic sides of Claude Code on this article, as a result of it’s a very good instrument that makes me far simpler as an engineer. Nevertheless, there are additionally some downsides to Claude Code.
The primary situation I’ve with Claude Code is that it’s a bit gradual typically, particularly after I evaluate it to utilizing Cursor. This occurs regardless that I’m utilizing Claude Sonnet 4.5 or Claude Opus 4.5 in Cursor.
I’m not completely positive why that is the case, however I typically discover that when analyzing log teams or performing easy fixes, Claude in Cursor is solely quicker than utilizing Claude Code.
One other draw back to Claude Code is that it’s purely terminal-based. If you wish to perceive your code higher and evaluation it, I might encourage you to make use of an IDE as an alternative. It also needs to be famous that there exists a Claude Code extension for VS Code, which may be put in in any VS Code fork as properly, akin to Cursor.
Conclusion
On this article, I’ve talked about Claude Code, protecting why and when you need to make the most of the instrument. I highlighted that Claude Code excels at coding whenever you don’t must carry out guide opinions of the code, which is increasingly typically the case with the fast enhancements in LLMs. I then lined particular strategies I exploit to get probably the most out of Claude, akin to heavy utilization of reminiscence and plan mode, in addition to attempting to use Claude Code to just about each drawback I encounter. I imagine coding brokers are right here to remain, and that you need to embrace the expertise as a lot as you’ll be able to. You need to continuously be testing the boundaries of what these coding agent instruments can do, to be able to grow to be as efficient an engineer as attainable.
👉 My Free Sources
🚀 10x Your Engineering with LLMs (Free 3-Day Email Course)
📚 Get my free Vision Language Models ebook
💻 My webinar on Vision Language Models
👉 Discover me on socials:
🧑💻 Get in touch
✍️ Medium
