Getting started with AI agents from scratch: installing, understanding GitHub, and running your first tool

This is an article that I’ll keep updating.

I open-sourced the tools I built and put them on GitHub. People have been telling me, “The stuff looks impressive, but I don’t even know what to do for step one.” Others asked, “What is GitHub? I click in and see a bunch of files, then what?”

So this post isn’t about introducing tools. It’s about walking you through getting your first one running.

If you’re already using Claude Code, skip straight to Section 4.

Last updated: 2026-07-24

Take 30 seconds to see if this post is for you

What you’ll achieve after reading: On your own computer, you toss a paper’s DOI to the AI, and ten-odd minutes later you get a critical appraisal report with citation checks and a level of evidence. Or you throw a PDF textbook at it and have it organize the book into note after note, complete with citations.

What you need to prepare:

• A computer (Mac or Windows both work, hospital/company computers are fine too, administrator privileges are not required)
• A paid Claude subscription, Pro is enough (about US$20/month). Free accounts don’t support Claude Code; this is a hard requirement.
• You don’t need to know how to code. Really. All you need is to know how to “copy and paste commands” and “not freak out when you see red error messages, just paste the whole thing back and ask the AI.” That’s it.

If you can do those two things, keep reading 👇

—

1. Same person, the tools just grew up

Let’s start with my own timeline, because I think this matters more than any tutorial:

• March: Brought AI onto my own computer for the first time. I did semantic search and automatically turned class recordings into notes. Back then, every time I did anything, I had to manually paste a “you can only answer based on the data I provide” prompt, watch it to make sure it wasn’t making things up, and click confirm at every step. It drove me crazy.
• April to May: Slowly ground book-converting, image-extracting, and OCR into a production line. Started turning repetitive tasks into single clicks.
• June: Set up a paper radar and a rehab course radar while prepping for the board examination. They would automatically push new stuff to me.
• July: Outpatient smart medical records, quota gatekeeping, and having two AIs review each other’s output. Started “managing a group of AIs.”

I didn’t get smarter in between. The tools just grew up, plus I learned to write every pitfall I stepped in into a rule.

I’m bringing this up to say: you right now are my March. And you’re luckier than I was back then. The tools you’re getting your hands on from day one are already the highly usable versions.

2. Claude Code is not a chatbot

This is where most people get stuck. The reason isn’t technical, it’s that the mindset hasn’t shifted.

The ChatGPT or Claude web version you usually use is someone who is “very good at talking.” You ask it a question, it answers you. It can’t see your files; you have to copy and paste to feed them to it.

Claude Code is different. It runs in your computer’s terminal, and it has hands:

• It can open files in your folders and read them.
• It can write new files and edit old ones.
• It can run programs, see the results, realize it broke, and go back and fix it itself.

How big is the difference? Take a real example: I tell it “turn chapter 12 of this textbook into notes,” and it will go split the PDF itself, call OCR if it finds scanned pages, extract images again if it grabbed them poorly, and finally write a dozen notes and save them into my Obsidian. I only said that one sentence the entire time.

This is the “agent” everyone is talking about. In plain English: you hand off a goal, it thinks up the steps itself, gets to work itself, and checks the results itself.

There’s another key point many people don’t know: Claude Code doesn’t only run in “code folders.” Whichever folder you open it in, it can read and write the stuff there. Your transcript folder, your patient education leaflet folder, any of them work.

Once this clicks, the way you talk to it will completely change. You won’t ask “what does this paragraph mean” anymore. You’ll just say “organize the stuff in this folder for me.”

3. Install in ten minutes, give yourself an early win

In this section, we’re not installing my tools yet. Let’s get you a quick win first, otherwise it’s easy to give up if you get stuck on setup early on.

Install (one command, really)

Good news: you don’t need to install Node.js, Git, or anything else first. One command handles it all automatically.

Windows: Search for “PowerShell” in the Start menu, open it, paste this line and press Enter:

irm https://claude.ai/install.ps1 | iex

If the left side of your screen says C:\> instead of PS C:\>, it means you opened CMD. Paste this instead: curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

Mac: Search for “Terminal” in Spotlight, open it, paste this line and press Enter:

curl -fsSL https://claude.ai/install.sh | bash

Check if it’s installed correctly

Type claude --version. If you see a version number like 2.x.x (Claude Code), you’re good.

Then type claude and press Enter. It’ll pop up your browser asking you to log in. Log in with your Claude Pro account, go back to the terminal, and once you see Login successful, you’re done. You won’t need to log in again.

Your first words

After installing, you’ll see a cursor waiting for your input. Don’t rush into big projects just yet. Find one of your own folders (with a few files in it), cd into it, and tell it:

What's in this folder? Check if there are any duplicate files for me.

It’ll go look itself, figure it out itself, and report back to you.

This moment is important. You’ll feel for the first time, “Oh, it actually got to work, it’s not just chatting with me.” Once you have that feeling, the rest comes easy.

Will it break my computer?

A friend told me: “Running an agent means you first have to acknowledge the risk of it deleting the wrong things, and usually you also need some command-line skills. People with those skills have already installed it themselves, and those without probably wouldn’t dare use it.” He’s right, so this section fixes both issues at once.

Let’s start with its built-in safeguards, which are actually smarter than you’d think:

• It knows how to distinguish what’s important. Operations like reading files or searching are just “taking a look,” and it does them without bothering you. When you do get asked, it means this step is going to change something, and that’s exactly when it’s worth your time to take a look. So “constantly being asked” isn’t as annoying as it sounds, and it acts as your seatbelt when you’re a beginner. Don’t rush to turn it off.
• Files it edits itself can be restored. It automatically saves a restore point before every edit. You can revert it by typing /rewind (or pressing Esc twice), like loading a game save.
• But restore points have limits: they only cover files it changed using its “edit tool.” If it deletes files using commands (rm, del), they won’t go to the recycle bin; they vanish instantly, and /rewind won’t save them.

That third point is something every tutorial should mention but rarely does. There are three layers of countermeasures:

• Use copies during your practice phase: create a practice folder, copy the files you want to process into it, and then get to work. By default, it treats the folder you opened it in as home turf. It’ll ask for your permission before touching anything outside (just agree, it’s not locked down).
• Turn your habits into its rules: just tell it straight up, “from now on, when deleting files for me, always move them to the recycle bin first, and save this in your rules file.” It’ll do it, and it’ll remember it in every conversation going forward (I’ll explain what the rules file is and how to grow it in the next post).
• Folders you’d cry over losing need backups: cloud sync or an external drive both work, or just tell it, “put this folder under git version control for me,” and let it buy the insurance for you itself.

💡 The one-sentence version: practice with copies, take a look when asked, route deletes to the recycle bin, and back up important stuff. Once you do these things, feel free to mess around safely. Even if something really goes wrong, you’ve only deleted a copy.

What if I get stuck (this is actually lesson one)

If it won’t install, or you see English error messages you don’t understand, don’t spend ages googling it yourself. Open the chat box on claude.ai and just ask like this:

“I’m a total beginner who has never used a terminal, I’m using Windows, and I want to install Claude Code. Please teach me step by step in the simplest way possible.”

If you’re stuck on an error, copy and paste the whole thing to it, and add “ELI5 (explain like I’m 5) what this means and what I should do.”

It’s your on-call engineer. This very act of “asking it when you don’t know” is the most important skill.

⚠️ Before you start playing, let’s draw one red line first: this is a cloud tool, what you type in will be sent to tech companies. For anything involving patient information, practice with fake or de-identified data first. Don’t put in a single API key or password. Build your intuition using fake data first, and only touch the real stuff when you can judge what’s okay to send and what isn’t.

When you do start letting it touch your work stuff, read this post next: When AI can do everything, I went back and double-checked the keys to my house. That was my own record of taking stock of my keys and tokens, the kind of check that is too late if you only do it after a disaster happens.

Learn in passing: understand what it’s doing, plus a few small commands

You barely ever have to type commands yourself (just tell it to do it), but it’s very much worth it to understand what it’s currently doing. These are the only ones that actually show up often in the terminal:

Just glance over the first five, no need to memorize them. The only reflex you need to train is: pause for half a second when you see rm.

Then there’s a batch of “slash commands” that you type directly to Claude Code itself. Typing / in the chat box will pop up a menu. Don’t memorize them, just recognize these few for now:

💡 The only hard stop in the subscription model is your quota. Once you start having the AI run longer tasks, you can check out claude-pacer, which I built. It turns /usage into a fuel gauge you can always see on your status bar. When it’s about to run out, it’ll automatically wrap things up and resume after the quota resets. Don’t bother installing it when you’re just starting out; just know it exists.

4. What is GitHub? What are you actually looking at when you click in?

My tools are all hosted on GitHub, so I need to spend a bit of time explaining it in this section.

Plain English definition

GitHub is a public library for storing code and tools, and it also conveniently records a complete history of “who changed which line and when” for everything.

For you, you just need to understand it as: everyone puts their finished tools up here publicly for free, and you can come grab them.

You don’t need to register for an account to download things. Registering is only for starring things or opening issue discussions.

On a page, you only need to look at four places

Click into any project (called a repository on GitHub, or repo for short), and you’ll see a bunch of files. Don’t panic, there are only four places to look at:

You can ignore all the other tabs (Actions, Insights, Pull requests…) for the foreseeable future.

How do you download things? Three methods

Method 1: Just tell Claude Code to go get it (highly recommended)

Open claude in your terminal and tell it:

Grab this repo https://github.com/drpwchen/paper-review-and-digest for me,
and install it following the instructions in its README.

Using an agent to install an agent’s tools—this is the correct way to use it 😆 Plus, it’ll read the README for you while it’s at it.

Method 2: Click Download ZIP on the webpage (most foolproof)

The green Code button → Download ZIP → unzip. Absolutely zero commands to learn.

Method 3: git clone (the standard way)

git clone https://github.com/drpwchen/paper-review-and-digest.git

If it says it can’t find git, then you’ll need to install Git first (just download the installer and click next). Or just use Method 1 or Method 2; the result is exactly the same.

How do you judge if a tool is worth using?

There’s a lot of stuff on the internet. I personally look at three things:

• Is the README clearly written: People who are too lazy to write a manual are usually too lazy to deal with your problems.
• Has it been updated recently (the last modified time is shown on the right side of the file list): Something untouched for two years is very likely incompatible with today’s environment.
• Does it have a LICENSE file: This is how the author authorizes you to use it. Mine are all MIT, which in plain English means “use it, change it, commercialize it however you want, I’m not responsible if something goes wrong.”

5. What is a skill? And what is MCP?

Okay, you know how to use Claude Code now, and you understand GitHub. So what exactly is that pile of stuff I put up there?

skill = Employee handbook

Imagine you hired a very smart newbie. He can do anything, but he doesn’t know the rules for how your unit does things.

He certainly knows how to “critically appraise a paper,” but he’ll do it his own way. One way this time, another way next time, and he’ll very likely miss important steps. What you want is: follow the same workflow every single time, check what needs checking, and cite what needs citing.

A skill is that employee handbook. It’s essentially a SKILL.md file, written in human language, saying “when you encounter this kind of task, do this for step one, do this for step two, and here are the situations where you stop and ask a human.” You drop it into the ~/.claude/skills/ folder, and Claude Code will read it itself when it’s time to use it.

So you’ll find that most of my repos aren’t “an app,” but rather a meticulously written workflow manual, plus a few small scripts that handle the math. Because the workflow itself is what’s truly valuable, not the code: it’s something I slowly figured out, finding out exactly where things break, only after reading hundreds of papers and organizing thousands of notes myself.

By the way, the design of skills is very clever: normally only the name and a one-line description are in front of him. He only loads the full content in when he actually uses it. So even if you put a massive reference manual inside a skill, it costs almost nothing when unused.

But that doesn’t mean you can install an infinite amount. The real cost lies in that one-line description: if you install dozens of them and the descriptions look similar, he’ll pick the wrong one, or fail to think of using it when he should. Just install what you’ll actually use, and write the description as “when you should call me.”

When the day comes that you feel “I want to turn a workflow of my own into a skill,” there’s an official plugin called skill-creator. Install it, then just tell it, “help me turn this workflow into a skill using skill-creator.” It will walk you through clarifying the steps, write the file, and even test it for you. This was how I got started early on myself. Later, you’ll realize you can even skip this and just state your needs directly and it’ll handle it, but having your hand held the very first time you write a workflow into a skill makes a huge difference.

MCP = External plug

A skill teaches him “how to do it”; MCP lets him “reach the stuff.”

MCP (Model Context Protocol) is a standard connector specification. Plug in an MCP server, and Claude gains a new ability: it can search PubMed, read your Gmail, or search your own note vault.

In a word: skills provide the workflow, MCP provides the interface.

⚠️ MCP used to have a bad reputation: every connected server would stuff the manuals for all its tools into every conversation you had. The more you turned on, the more expensive it got. This has been fixed now. Claude Code defaults to using tool search, and tool specs only load when needed.

So the criteria for choosing an MCP is no longer about saving tokens, it’s about what it touches: it connects to your email, your files, your accounts. Ask “what of mine can this server access” before installing it. That is far more important than saving tokens.

6. Hands-on: install your first tool

Let’s use paper-review-and-digest as an example. The reason is it has the lowest barrier to entry. No extra stuff to install, no graphics card needed, just toss it a DOI and you get output.

It contains two skills inside:

• /paper-review: Can I trust this paper? Journal credibility, author track record, verifying reference authenticity, risk of bias, and level of evidence (GRADE).
• /paper-digest: What is this paper talking about? Reformatting the whole thing into easy-to-absorb teaching notes, with self-test cards at the end.

Step 1: Tell Claude Code to install it for you

Don’t type the commands yourself. Open claude in the terminal and paste this entire block to it:

Help me install https://github.com/drpwchen/paper-review-and-digest
These are two Claude Code skills. Please install it according to its README:
Put paper-review/ and paper-digest/ into my skills folder,
copy config.example.yaml to config.yaml,
and when you're done, tell me if there's anything else I need to fill in myself.

It will go grab it itself, determine which path it should go to on your operating system itself, and copy the config file itself. All you have to do is hit agree when it asks “is this okay?”.

💡 This isn’t being lazy, this is the correct way to use it. Manually typing paths is the step most prone to errors (Mac and Windows format them differently, the folder might not even exist yet), and he won’t make typos; plus, he can read the English in the README that you don’t understand.

By the way, this is also your general rule for installing any tools later on: throw the repo URL at him, and tell him to read the README and install it.

Step 2: Check the config file

After installing, he’ll probably tell you where the config file is. You can just ask him: “Do I need to fill anything in this config?”

The answer is usually: nothing to fill. The note vault path, library full-text access, review card websites—these are all optional. Leaving them blank just turns them off; the core critical appraisal features run just fine.

Step 3: Run it

Restart Claude Code and type:

/paper-review 10.1002/pmrj.70039

This is a systematic review on neuromuscular training for knee osteoarthritis, published in physical medicine and rehabilitation (PM&R)’s very own journal, PM&R.

Then just watch him work. He’ll go check the journal, check the authors, verify one by one whether the references actually exist, determine which risk of bias tool to use for this study design, and finally calculate the level of evidence.

I actually ran this paper myself. The tool ultimately calculated the level of evidence as Very Low, which matches what the authors reported themselves in the article.

And that level isn’t asking the model “what level do you think it is.” The model is only responsible for judging how severe each of the five domains is, and finally a script recalculates it from scratch according to GRADE rules. The DOIs in the references are also legitimately sent to CrossRef via API to confirm they exist. Because it’s a language model, not a calculator. The model judges, the script does the math. Dividing the labor this way is safe.

The first time you watch him finish a run, you’ll probably understand why I spent time building these things.

7. The tools I open-sourced: two production lines

My tools aren’t scattered pieces. They form two production lines plus a few peripherals:

📚 The paper production line

📝 The note production line

So where exactly should I start?

• People who read papers → paper-review-and-digest, fastest payoff.
• People with a ton of PDF textbooks preparing for exams → textbook-to-note.
• People already using Obsidian who have so many notes they can’t find anything → vault-search.
• People who just want to play around and see what agents can do → ytscribe, fetching YouTube channel transcripts using speech recognition (ASR); totally unrelated to medicine, anyone can try it.

📦 Want to see the complete list of tools (not just ones I wrote myself, but also ASR, OCR, free hosting, home servers, etc.)? Go to this post: My complete AI toolbox

8. What’s next

You can install and run things now. The next step is using them well: that’s an entirely different beast. I wrote another post about it:

👉 How I talk to AI agents: Three mental models, why the old ways of prompting are obsolete now, and how to write the pitfalls you step into into rules so it stops making those mistakes.

If you get stuck:

• Problems with the tools themselves → Just go open an Issue on that repo, I’ll see it.
• General usage questions → Leaving a comment under my Facebook post is fine.
• Want to ask in more detail, or want to see experiments I haven’t finished yet → I have a subscriber community, and I prioritize answering there (but no personal medical advice, no exceptions).

Finally, as I always say: You don’t need to know how to do it first, you just need to know how to ask.

These tools are all things I use myself every single day. I open-sourced them because I felt they’d be useful to others; they’re not products. Taking them and modifying them to suit your own needs—that is their true intended use.

If you get stuck anywhere, or if you modify it into a better version, you’re more than welcome to let me know~ 😊

—

This article will be continuously updated. If a tool gets a major overhaul, or I step into a new pitfall, I’ll come back and add to it.