01 / BEFORE YOU BEGIN
VS Code is a playground for working with AI. New to VS Code? Learn what it is, then prepare your lab step by step.
What is VS Code? VS Code (Visual Studio Code) is an application for opening and working with files on your computer. Connect AI and other tools to view research materials, ask AI for help and inspect results in one place.
Imagine setting up a playground VS Code is the playground . Invite Claude to join you. A basketball enables basketball, and a football enables football: similarly, the files and tools you provide and the tasks you request determine what happens here.
Playground → VS Code A workspace bringing files, AI and tools together.
Your teammate → Claude Reads your requests and helps with research tasks such as analysis and writing.
Balls and equipment → Files and tools Provide protocols, data and papers, and connect tools such as Python and R for the work you need.
How this applies to research Provide a research protocol and ask: “Please review this study's design.”Prepare data and analysis tools and ask: “Please summarize participant characteristics in a table.”Provide the protocol and analysis results and ask: “Please propose a manuscript plan.”
Let's prepare the playground Install VS Code, connect AI and tools, then prepare a folder for research materials. That folder is your lab's workspace . You can follow from the beginning without completing Basic.
Your computer Install → Lab → Project → Research Install VS Code and tools, then open the team folder. Prepare a dashboard and project before analysis and writing. Finally, learn to use the terminal and manage multiple projects.
Check your account too This guide uses Claude. Prepare a Claude account and check Claude Code access requirements and usage limits after signing in. Claude account guide ↗
Already completed Basic? Keep using your existing research-team folder and research files. Check tools already installed and move on. Do not download the team again and overwrite your working folder.
Your VS Code research workspace at a glance Open research files, request work from AI and review results on one screen.
Actual Mac workspace · Select the image to enlarge it. ① Left · Research files Find folders and files inside research-team. Each study's protocol, data and results are collected under projects.
② Center · Claude conversation Enter a task in the input box below and review the AI's plans and results.
③ Right · Open document Read the file selected on the left. This screenshot shows the writing researcher's role instructions.
This example shows the workspace after opening the team folder. You do not need to recreate it yet. The next steps cover installation and folder setup. Arrange the screen to suit your workflow.
What to check in this step You have identified your computer and account and understand what to install and where research will take place.
The Next button moves through the guide. It does not automatically verify installations or execution results. 02 / INSTALL VS CODE AND EXTENSIONS
Prepare your workspace and four tools. Extensions add tools inside VS Code. Install all four listed below.
Official VS Code download ↗
Windows Run the Windows installer, complete installation and open VS Code.
Mac Open the download and move Visual Studio Code to Applications, then launch it. Extract the ZIP first if necessary.
1. Windows · Install and launch VS Code Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Windows recording · 39 sec · Zoomed selections and Korean captions · Original audio retained
Mac · Download and install VS Code Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Mac recording · About 37 sec · Download waiting time removed · Zooms and Korean captions · Silent
Already installed? Open your existing app The video opens the download and drags Visual Studio Code into Applications. If warned that the app already exists, it cancels. For a first installation, finish copying and launch Visual Studio Code from Applications. The video ends at the copy dialog; step 03 shows VS Code running.
1 Open Extensions Select the blocks-shaped Extensions icon on the left. You can also use Ctrl+Shift+X on Windows or Command+Shift+X on Mac.
2 Check the name and publisher, then select Install Find each extension using the links below or the search box. Use the listed extension ID to distinguish similarly named items.
3 Check that all four are installed Review the installed list. Select Reload if asked to reload VS Code.
Install R itself separately in the next step R Extension Pack alone cannot run R calculations. Documents unsupported by Office Viewer can be opened in Word, Excel or another appropriate app.
2. Windows · Install extensions Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Windows recording · 1 min 50 sec · Zoomed selections and Korean captions · Original audio retained
Install the two items not shown in the video too The existing video shows Claude Code, Codex, Python, an individual R extension and Office Viewer. Skip the Codex installation. R Extension Pack and Live Server must also be installed from the list above. An individual R extension is not the same as R Extension Pack. Even if you install the Python extension, check the Python interpreter itself in step 04.
What to check in this step VS Code opens and all four extensions appear in the installed list.
The Next button moves through the guide. It does not automatically verify installations or execution results. 03 / SIGN IN AND CONNECT THE AI
Open an AI conversation inside VS Code. In this step, open Claude and sign in. Actual research requests begin after preparing the team folder.
1 Open Claude Code Select the Claude icon on the left. If this is your first use, follow the sign-in instructions to connect your Claude account.
2 Check the connection Confirm that the Claude conversation opens and accepts text input.
Video guide: open the Claude conversation Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Mac recording · About 34 sec · Recording-stop screen removed · Highlighted selections and Korean captions · Silent
Only open the conversation at this stage Select Claude on the left and sign in. The existing video also opens Codex; follow only the Claude section for this guide. Codex installation and sign-in are unnecessary. Prepare the team folder in step 05.
Official Claude Code guide ↗
What to check in this step The Claude conversation opens. You have not yet connected the team folder or started a research request.
The Next button moves through the guide. It does not automatically verify installations or execution results. 04 / INSTALL PYTHON AND R
Install the tools that execute your analyses. Follow the Python installation video and install R too. Fully quit and reopen VS Code afterward.
1. Install on Windows About 2 min 11 sec · Installation waiting time is shown at 6× speed.
Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Open the official Python downloads page ↗ Download and install the official Python Install Manager. Open a new PowerShell window and enter python --version. If Python installation starts, wait for it to finish. Run python -c "print('PYTHON_OK')" and check that PYTHON_OK appears. Also run python -m pip --version to check that package-installation support is available. 1. Install on Mac About 1 min 43 sec · The Spotlight shortcut is Command (⌘) + Space.
Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Open the official Python downloads page ↗ Download and install the official Python 3 macOS installer (.pkg). Press Command (⌘) + Space to open Spotlight, then search for Terminal. Run python3 --version, then python3 -c "print('PYTHON_OK')". Check the version and PYTHON_OK output. Also check python3 -m pip --version. Additional guidance beyond the video · Internet certificates For Python installed from python.org on Mac, prepare the additional certificates described in the official guide. They help Python verify websites when downloading data. This is separate from testing a basic calculation.
In Finder → Applications → Python 3.x, locate Install Certificates.command and double-click it. Python installed by other methods may not include this file.
2. Install R Official R download ↗ 1 + 1. Check that the result is 2.
Video guide: download R and check execution · Mac Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Mac recording · About 41 sec · Zoomed downloads and input · Korean captions · Original audio retained
For a first installation, open the downloaded .pkg file and finish installing, then launch R. Installation itself is omitted in the video. Open Spotlight with Command + Space .
3. Check actual execution in Claude inside VS Code After installation, reopen VS Code and paste the request below into Claude. There is no need to create a separate check.R file. The video already has research-team open, but this step only checks Python and R execution. Prepare the team folder in the next step.
Video guide: check Python and R installation Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Mac recording · About 47 sec · Zoomed questions and results · Korean captions · Original audio retained
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Python and R installation and execution check Copy prompt
Results differ between computers Your versions and installation paths need not match the video. Check actual calculations: 2 + 3 = 5 in Python and 1 + 1 = 2 in R. Prepare additional analysis tools as each study requires. Do not act on the video's project-creation suggestion yet; move to the next step.
What to check in this step You have checked actual calculations in Python and R. If something does not run, give Claude the error and resolve it before continuing.
The Next button moves through the guide. It does not automatically verify installations or execution results. 05 / PREPARE THE TEAM FOLDER
This folder is your lab. Extract the package in a convenient location on your computer and open that folder in VS Code.
What the team setup package contains It includes shared AI instructions, researcher roles and skills, dashboard tools, and a sample protocol and data. Keep them together in research-team: this folder becomes your lab's workspace.
Download the research-team setup package (.zip) Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Download and extraction guide · Explanatory video 1 Extract the ZIP On Windows, right-click the ZIP and select Extract All. On Mac, double-click it. Confirm that the extracted research-team folder contains START_HERE.md, sample and projects.
2 Open the lab in VS Code Select research-team using File → Open Folder. If asked whether to trust the files, first confirm that this is the downloaded folder whose contents you have checked.
3 Ask AI to check readiness Paste the prompt below into the AI conversation in the current window. Do not create a project or begin research yet.
research-team/ ← Folder to open in VS Code
├─ START_HERE.md Getting-started guide
├─ team/ Shared team materials
├─ guides/ Usage guides
├─ sample/ Sample protocol and data
└─ projects/ Workspaces for your studies Video guide: open and check research-team Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Mac recording · About 2 min 50 sec · From folder selection to Claude's review · Full screen preserved · Korean captions
Check the team folder after opening it The video opens research-team, examines a protocol, and opens both Claude and Codex. This guide uses only Claude; skip the Codex part. It then asks Claude to review the working folder. The findings concern the folder in the video; your results depend on the downloaded package and computer environment.
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Lab readiness check request Copy prompt
What to check in this step You have opened research-team and received checks of shared instructions and the analysis environment.
The Next button moves through the guide. It does not automatically verify installations or execution results. 06 / UNDERSTAND YOUR TEAM
Folders define roles. Records keep research moving. The AI lab has the same structure and working methods as Basic.
The folder you open determines the conversation's role Conversation in research-team = lab coordinator
Conversation in projects/project01 = that study's team lead
Conversations do not automatically share memory. Each team saves progress records, and the coordinator reads those records to understand its status.
Who works with you?
Eight roles. Decisions stay with the principal investigator. Principal investigator — you / Team lead — the AI in your current conversation You set the research direction and make important decisions. The team lead distributes tasks to the researchers it needs and summarizes their results and reviews. You do not need eight open windows.
01 / EVIDENCE Literature researcher Develops search strategies and evidence tables and collects verified papers.
02 / PLANNING Study-design researcher Prepares questions, variables, analysis plans and protocol drafts.
03 / DATA Data-management researcher Asks once whether data are anonymized, then examines the codebook and data quality.
04 / ANALYSIS Statistical researcher Works on appropriate methods, code, effect sizes and 95% confidence intervals.
05 / MANUSCRIPT Writing researcher Connects actual results with evidence to draft the manuscript.
06 / QUALITY Review researcher Separately examines study design, numbers, interpretation and reporting guidelines.
07 / REFERENCES Reference-verification researcher Checks bibliographic details separately from whether a paper supports the claim.
08 / CONTINUITY Records researcher Documents decisions, unresolved issues and next tasks.
Check whether your app can invoke separate role-based AI researchers. If it cannot, drafting and self-review are possible, but do not label them as completed independent review.
What can you ask the team to do?
Start with the work you need and return to earlier tasks when needed. Questions and literature What do you want to know? Design and planning Variables and analysis plan Data preparation Confirm anonymization Analysis Select the analyses needed Writing and review Manuscript, numbers and reporting guidelines Records and resuming work Decisions and next tasks These are not mandatory sequential stages. Choose tasks based on your study type and available materials. “Clinical Trial” in the illustration is an example name; observational studies and systematic reviews use their own appropriate setup.
Built on public projects: Harness · Minho Hwang (role configuration), MedSci Skills · Yoojin Nam (research skills), Clinical Research Harness (research record principles), CoVe (reference verification) · Ki-Hyun Jeon. Source-specific terms are included in the distributed folder.
Team working principles Do not invent results or references. Start research where needed and record changes and the work actually performed. Distinguish drafting from separate review and clearly mark anything unverified. Ask once, “Have patient identifiers been anonymized?”, before working with the data. Do not repeatedly ask about the same data or require anonymization certificates. Completing earlier stages or locking a plan is not a prerequisite for starting.
To continue the same study another day At the end, ask: “Record today's work, decisions and next tasks.” Next time, open the same project folder and say: “Read the previous records and continue.”
You do not have to switch to Advance simply because you have multiple studies. Basic and Advance offer different working environments.
What to check in this step You understand the distinction between lab coordination and individual projects and why researcher instructions and records are needed.
The Next button moves through the guide. It does not automatically verify installations or execution results. 07 / SET UP THE DASHBOARD
See your previously invisible lab on screen. Make this request in the VS Code window with research-team open. Unassigned teams and an empty pipeline are normal when you have no projects yet.
This example already has a project registered. Your initial dashboard has no studies. Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Dashboard creation request Copy prompt
1 Find dashboard.html Locate dashboard.html at the path reported by the AI.
2 Open with Live Server Right-click the file and select Open with Live Server. Check that your lab dashboard opens in the browser.
3 Adjust display settings Change the name, photo, light/dark mode, text size and spacing. Live Server opens the page; it does not update the research records themselves.
Video guide: create and personalize the dashboard Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Mac recording · About 39 sec · Zooms and Korean captions · Waiting time at 5× speed · Recording-stop screen removed · Silent
After creating the dashboard, explore the teams and researchers, then use Display settings at the top right to choose and save Dark mode. Zero studies and an empty pipeline are normal without any projects.
What to check in this step Your local dashboard opens. It shows the latest research records rather than monitoring live activity.
The Next button moves through the guide. It does not automatically verify installations or execution results. 08 / PREPARE A PROJECT
Create one folder per study. On your computer, create project01 under projects and add the sample folder's files. Keep using the research-team window until team connections are ready.
Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Create project01 and copy sample files · Actual Mac recording 1 Create projects/project01 Use File Explorer, Finder or VS Code's file list.
2 Add the protocol and data Copy GLP1_AMI_RCT_protocol.docx and GLP1_AMI_RCT_data.xlsx from sample into project01.
3 Connect the team from the lab coordination window Send the request below to the AI in the window with research-team open. Having two files in a folder is different from having the team instructions connected.
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
project01 team setup request Copy prompt
You do not write a new START_HERE.md yourself Prepare the shared lab once. Each project receives instructions referring to the common team, a starting guide and progress records. Continue after the AI verifies the created files and connections and reports setup complete.
Video guide: run the project setup prompt Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Mac recording · About 40 sec · Zoomed input and completion report · Waiting time at 6× speed · Korean captions · Silent
What to check in this step project01 contains both files, and you have confirmed its instructions, shared-team connections and progress records.
The Next button moves through the guide. It does not automatically verify installations or execution results. 09 / PRACTICE
Start manuscript work in a new project window. Keep the lab coordination window open and open another VS Code window for project01.
1 File → New Window In the new window, select File → Open Folder → research-team/projects/project01.
2 Open the AI conversation in that window Check that the top-level folder on the left is project01. Use the prompt below to read the prepared team instructions and materials.
3 Review and approve the manuscript plan Review the proposed research question, analyses, tables, figures and manuscript plan. Approve it or request changes.
4 Check the actual outputs Open the tables, figures and manuscript. Distinguish outputs from sample data from actual clinical research findings.
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Manuscript planning and execution request Copy prompt
Video guide: open a new project window and begin practice Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Mac recording · About 1 min 42 sec · Zoomed key screens and Korean captions · Waiting time shortened · Original audio retained
In the video, GLP1_AMI_RCT is the example folder corresponding to project01 in this guide. Follow along using your own prepared project folder. The practice reviews a draft, tables and figures created from synthetic data; the researcher must check all content and numbers.
What to check in this step You have reviewed the plan in the project window and checked actual outputs and review points for the approved work.
The Next button moves through the guide. It does not automatically verify installations or execution results. 10 / CREATE YOUR OWN PROJECT
Start with what you have now. A question is enough, or you may already have a protocol and data. Choose whichever of the five requests below you need now.
Prepare your own project the same way In the research-team window, ask for a study folder under projects and connections to the shared team. Put your materials there and confirm the connections, then open that project folder in a new VS Code window. Begin with the work each project needs.
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Your own project setup request Copy prompt
1. Ask a research question Adapt the examples and submit them in your own project window.
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Ask a research question Copy prompt
2. Share a research hypothesis Adapt the examples and submit them in your own project window.
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Share a research hypothesis Copy prompt
3. Develop a research protocol together Adapt the examples and submit them in your own project window.
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Develop a research protocol together Copy prompt
4. Plan an analysis with your data Adapt the examples and submit them in your own project window.
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Plan an analysis with your data Copy prompt
5. Draft a manuscript from data and a protocol Adapt the examples and submit them in your own project window.
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Draft a manuscript from data and a protocol Copy prompt
Video guide: work on multiple studies in separate windows Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Mac recording · About 1 min 5 sec · Zoomed menus and folder selections · Window arrangement at 2× speed · Korean captions · Silent
File → New Window adds a window. Open a different research folder in each window. Keep a lab coordination window open and request tasks separately in individual project windows. The study names and files shown are real usage examples; you do not need to reproduce them.
Opening windows alone does not start research. Request work in each window. Conversations do not share memory automatically, so save progress records and avoid editing the same file in multiple windows at once.
What to check in this step You have requested the work needed in your own project window and identified the next task.
The Next button moves through the guide. It does not automatically verify installations or execution results. 11 / TERMINAL AND MULTIPLE PROJECTS
Work more flexibly in the same lab. The terminal is a command-entry panel at the bottom of VS Code. The extension conversation and the terminal AI have different installation steps.
Why install the standalone CLI? The CLI runs Claude through commands. In addition to entering requests yourself, you can let another program call Claude, assign a task, and receive the result.
Extend your lab with background work and automation For example, you can build a workflow that asks a literature agent for a summary, passes it to a review agent, and collects the results into a report. The CLI provides a way to run and connect these tasks.
You can also use it to request a scheduled status summary or dashboard update. Scheduling and result handoffs require separate setup, and the computer and required execution environment must be running.
Do steps 1–10 require a separate CLI installation? Steps 1–10 in this guide use the Claude extension in VS Code. The extension includes its own runtime and supports subagents. Install the standalone CLI to run the claude command in a terminal or connect it to external automation.
Does installing it make agents communicate automatically? Installing the CLI does not automatically share memories between conversations or make your lab exchange tasks. You still need to define responsibilities and specify which results go to the next agent. In this lab, work continues by reading project instructions and progress records.
This step covers installing and running the CLI. It does not set up scheduled jobs or automatic handoffs between agents.
Official documentation: VS Code extension and CLI · Run Claude programmatically
1 In the project folder, select Terminal → New Terminal Use pwd or Get-Location on Windows, or pwd on Mac, to confirm that the current location is your intended project.
2 Install the terminal AI Follow the official instructions below for your operating system, then open a new terminal. Installing the extension does not automatically prepare the terminal command.
3 Run and sign in Check with claude --version, then enter claude. Follow the first-time sign-in instructions.
4 Read records to resume research A new conversation does not share the previous window's memory. Ask it to read project instructions and progress records and report the next task.
Install Claude Code CLI Official installation guide ↗ irm https://claude.ai/install.ps1 | iexcurl -fsSL https://claude.ai/install.sh | bash
Video guide: install Claude Code CLI · Windows The black window on the right is PowerShell, where you enter commands. On your keyboard, use Windows key → Search for PowerShell → Open Windows PowerShell to launch it. If Windows Terminal opens, use its top ∨ menu → Windows PowerShell option. The input prompt begins with PS , which helps identify it. Your window color may differ.
Original video retained: interface text and captions are in Korean. Follow the English instructions on this page; folder names and commands are unchanged.
Actual Windows recording · About 1 min 6 sec · Zoomed PowerShell and Korean captions · Waiting time shortened · Original audio retained
The left window is Claude on the web for reference; enter installation commands in PowerShell on the right. The PATH adjustment illustrates how to address a warning. Sign-in and repeated troubleshooting are omitted. On first launch, follow the browser instructions to sign in with your account.
This video demonstrates Claude Code for Windows installation. After checking installation, run it in your own project folder as described below. Installation commands and troubleshooting are also available in the official Claude Code installation guide .
Window 1 · research-team Overall status and project setup
Window 2 · projects/project01 Sample study analysis and writing
Window 3 · projects/my-study Your own study
Give each window a clear responsibility Use separate windows for different projects. When reviewing one project in multiple windows, assign writing and review separately and avoid simultaneous edits to the same file. The coordinator reads project records to summarize progress.
Copy and paste into the AI conversation in VS Code Role → Goal → Inputs → Tasks → Report format → Completion criteria
Resume research in a new conversation Copy prompt
For opening multiple research windows, see the previous step's video, “Work on multiple studies in separate windows” .
What to check in this step You have checked that AI runs in the terminal and learned to continue work using project-specific windows and records.
The Next button moves through the guide. It does not automatically verify installations or execution results.