DEV Community: Qwerty

Automatically hide _assets folders in Obsidian (until you need them)

Qwerty — Tue, 31 Mar 2026 01:28:00 +0000

This used to be you (and by you I mean me).

Articles/
  Some Guide/
    _Some Guide.md
    Another take.png
    Final result.png
    How to do that.png
    How to do this.png
    ...
  Another Article/
    _Another Article.md
    Screenshot 1.png
    Screenshot 2.png
    ...

This is you with the Custom Attachment Location plugin.

Articles/
  > Article A_assets/
  v Article B_assets/
      file-1.png
      file-2.png
      ...
  > Article C_assets/
    Article A.md
    Article B.md
    Article C.md

It works, it keeps things tidy, but after a while the file explorer gets noisy anyway.

So I looked around.. , and around.. , and found that it was pretty common wish on the Obsidian Forums, one such dated back to 2021.

After digging around for a bit more I found a stupidly simple brute-force CSS snippet....

/* hide path's with __before and ending with__ */
.nav-folder-title[data-path^="__"][data-path$="__"] {
  display: none;
}

🤯🤯🤯🤯💡💡💡 and it hit me.

This article is my take on this CSS-only "hack" that you can drop straight into Obsidian > Appearance > CSS snippets without installing anything.

What this snippet does

Shrinks every folder whose name ends with _assets (configurable)
Expands the _assets folder only for the currently active folder (when one of its notes or its assets is focused/active)
Includes optional css for scrolling within assets.
Includes an optional “warning” highlight to identify folders that are expanded but visually collapsed (more on why below)

Note: In the gif below, the two visible collapsed assets folders are displayed as a thin gray line (customizable).

"Install" (Obsidian CSS snippet)

Recommended: Use the CSS Editor plugin.

With the editor:

Use the CSS Editor from the sidebar Ribbon to update an existing css snippet or create a new one, e.g. hide-assets-folders.css
Paste the CSS code below into it
It should already be enabled. You can check by using the CSS Editor button again.

Without the editor:

Copy the snippet into a file, e.g. hide-assets-folders.css
Place it in your vault’s snippets folder:
- Settings → Appearance → CSS snippets → (open snippets folder)
Enable it in Settings → Appearance → CSS snippets

Recommended: Use the Custom Attachment Location plugin to keep your attachments organized in _assets folders next to their notes.

configure "Location for new attachments": ./${noteFileName}_assets

The CSS snippet

/* Hide assets folders – https://forum.obsidian.md/t/automatically-hide-assets-folder-when-not-focused/111617 */
.nav-folder:has(>[data-path$='_assets']) {
  /* Base style for every assets folder (visible / hidden). */
  background: color-mix(in srgb, currentColor 5%, transparent); /* Highlight all assets folders. */
  max-height: 4px; /* The height of the hidden assets folder. */
  overflow: hidden; /* Hide everything beyond the max-height. */
  transition: max-height 500ms ease-in-out; /* Enable smooth expand/shrink animation. */

  /* Style overrides for active folders (has an .is-active child). */
  .nav-folder-children:has(>*>.is-active) > &,
  .nav-folder-children > &:has(.is-active) /* When the assets folder itself has an active asset. */
  {
    background: color-mix(in srgb, currentColor 3%, transparent); /* Highlight for active expanded assets folders. */
    max-height: 150px; /* Some large number is needed for the smooth effect. Too large ruins the animation, too small may limit assets visibility. Set at least 25px larger than the scrollable area below. */

    /* Optional: Enable scroll within assets folders. */
    .tree-item-children {
      max-height: 125px; /* Max height of the scrollable area. */
      overflow-y: auto;
    }
  }
}

/* Optional: Highlight forgotten NOT-collapsed assets folders that break menu items rendering due to expecting their actual height. */
.nav-folder-children:not(:has(>*>.is-active)) > .nav-folder:not(.is-collapsed):has(>[data-path$='_assets']):not(:has(.is-active)) {
  background: color-mix(in srgb, red 30%, transparent);
  max-height: 10px; /* Make it slightly larger so that it can be clicked to hide comfortably. */
  * { color: transparent }
}

How it works (quickly)

It targets folders whose data-path ends in _assets using:
- :has(>[data-path$='_assets'])
“Hidden” is implemented by collapsing the folder’s height:
- max-height: 4px; overflow: hidden;
When a note within that folder becomes active (.is-active), the child _assets folder becomes “tall” again:
- max-height: 1000px;

Obsidian’s file explorer DOM structure makes :has(...) a surprisingly powerful tool for this kind of UI behavior.

Known drawbacks (and why they happen)

1) If there are multiple `_assets` folders in the active folder…

With pure CSS, you can’t easily say:

“Show only the _assets folder that corresponds to the active note.”

CSS can only express relationships like “some descendant/sibling is active”, not “the descendant matches this sibling by name”.

So if your active folder contains multiple _assets folders, the logic can’t reliably reveal just one of them based on the current note.

Possible fix: a small JS snippet (via something like JavaScript Init) or a tiny plugin that toggles a class on the “correct” assets folder.

2) “Hidden” folders still affect virtualization / scrolling

This one is subtle.

Even though the folders are visually reduced, the file explorer’s rendering system still “counts” their size as if they were normally present. This is especially problematic when they’re expanded but visually limited by CSS. Obsidian virtualizes long lists (it dynamically renders/unrenders items based on scroll). When invisible elements inflate the expected height, rendering can happen too early/late, and you can end up with odd behavior like:

items being unrendered sooner than expected
the active note being pushed out of the rendered range
the _assets folder disappearing again because the active item DOM isn’t present anymore

That’s why the snippet includes a “warning” rule that highlights in red _assets folders that are not collapsed (expanded state) but currently hidden. In practice, it’s a signal to collapse them for smoother explorer behavior.

Recommended usage pattern

Use a consistent naming scheme like: SomeNote_assets/ or Folder_assets/
Keep _assets folders collapsed by default
Let the snippet reveal them only when you’re inside that folder
If you run into the virtualization glitch often, consider:
- manually collapsing assets folders
- or moving to a JS-based approach that truly removes/hides nodes rather than just shrinking them visually

What I might build next

If there’s interest (or if it annoys me enough), the next step is a tiny JS helper/plugin that:

identifies the active note
derives the “matching” _assets folder name
toggles a class on only that folder
optionally collapses other _assets folders automatically to prevent the virtualization issues

Credits / inspiration

This is based on and inspired by the forum CSS approach. Originally introduced in this thread.

Delightful TypeScript Mock Generator - Generate Realistic Test Data from Your Interfaces

Qwerty — Wed, 14 Jan 2026 02:57:31 +0000

Hello fellow developers! 👋

I was frustrated developing with a single example json file, so I called to the internet, but the internet didn't call back. The only thing I found were other lost souls looking for answers... so I made my own and I find it incredibly delightful to play with. I wanted to share it with you, because I think you might like it too.

It's a TypeScript Mock Data Generator that creates realistic test data from your interfaces or other JSONs.

Instead of manually writing mock data or struggling with generic generators, just paste your TypeScript interface or existing JSON example and get properly structured, realistic data instantly.

🔗 Try it live: https://ts-mock-generator.qwerty.art/

🔗 GitHub: https://github.com/ackvf/ts-mock-generator

Why I Built This

I was tired of:

Waiting for backend APIs and working with single static example
Manually creating mock data for development or tests
Getting { name: "string", email: "string" } from basic generators
Losing time on tedious setup for prototypes
Creating inconsistent test data across projects

So I made a tool that just... works. Copy, click, paste, BAM!, click and you're done. Simple as that.

The Cool Parts

1. Smart Field Recognition

The generator actually reads your field names and generates appropriate data. No configuration needed:

interface User {
  email: string       // → "alex.chen@example.com"
  firstName: string   // → "Alex"
  age: number         // → 32
  city: string        // → "San Francisco"
  phoneNumber: string // → "+1-555-0123"
  avatar: string      // → "https://cloudflare-ipfs.com/ipfs/..."
}

It recognizes 50+ common patterns: emails, names, addresses, phone numbers, UUIDs, URLs, dates, prices, and more.

2. Type Names Matter (The Secret Sauce!)

Can't change property names? Name your types meaningfully, and the generator picks up on it:

type First = string
type Age = number
type Email = string

interface Person { // Uses the type name as a hint!
  n: First         // → "Emma" 
  a: Age           // → 27
  c: Email         // → "emma@example.com"
}

3. "Invalid" Patterns That Work Beautifully

The tool allows some TypeScript patterns that are technically invalid but create amazing mocks:

type First = string
type Age = string

interface Data {
  // This "breaks" TypeScript index signature rules...
  users: { [name: First]: Age }
} 
  // But generates perfect output:
  // { "Emma": 24, "James": 19, "Sofia": 31 }

This is technically breaking TypeScript rules (type aliases should be transparent), but for mock generation? It's perfect! The generator treats type names as semantic hints.

The generator tolerates these "violations" because they make for better, more readable mock data!

4. Reproducible Results

Set a seed number, get the same data every time:

Seed: 42

Your last 3 seeds are saved and clickable—super handy when you finally clicked enough times to get the perfect random mock, but you clicked one too many!

5. Advanced TypeScript Support

It handles the complex stuff too:

Template Literals:

type Category = 'ELEC' | 'FURN' | 'CLTH'
type Size = 'SM' | 'MD' | 'LG'
type SKU = `${Category}-${number}-${Size}`
// Generates: "ELEC-8472-LG"

Insane depth is not a problem!

type Digit = '0' | '1' | '2' | '3' | '4' | '5' | '6' | '7' | '8' | '9'
type FourDigit = `${Digit}${Digit}${Digit}${Digit}`
type CardNumber = `${FourDigit} ${FourDigit} ${FourDigit} ${FourDigit}`

interface Card {
  number: CardNumber // → "8452 4675 4126 6009",
}

Utility Types:

interface Config {
  partial: Partial<User>
  picked: Pick<User, 'name' | 'email'>
  omitted: Omit<User, 'password'>
  lowercase: Lowercase<string>  // → "john"
}

Intersection Types:

interface Employee {
  info: { name: string } & { id: number } & { role: string }
  // Merges into: { name: "John", id: 123, role: "Engineer" }
}

Collections:

interface Store {
  inventory: Map<string, number>
  categories: Set<string>
  tags: [string, string, string]
  employees: {
    [name: string]: Age // → "Elbert Funk": 63
  }
}

6. JSON ↔ TypeScript

Works both ways:

TypeScript → JSON: Generate mock data
JSON → TypeScript: Paste API response, get interfaces

Great for working with external APIs!

Real Usage Examples

Quick prototyping with Realtime output:

The generator updates mock data as you type, so you can quickly iterate on your interfaces and see realistic data instantly. It will not break on errors, just shows nulls until valid.

interface BlogPost {
  id: string                            // → "b3f1c2e4-5d6a-7b8c-9d0e-f1a2b3c4d5e6"
  title: string                         // → "Understanding TypeScript Generics"
  author: string                        // → "Sophia Lee"  
  content: stri| // <- still typing     // → null
  tags: string[]                        // → ["typescript", "programming", "webdev"]
}

Test data:

Always use seed 12345 in tests
Everyone gets identical mock data
Tests are reproducible

Database seeding:

Generate 1000 users
Seed: 999
Same data every dev environment reset

Under the Hood

Built with:

Next.js for the UI
TypeScript Compiler API for parsing
Faker.js for realistic data generation
Runs 100% in your browser—no backend!

About Faker.js

The realistic data comes from Faker.js, which has tons of generators for names, addresses, images, lorem ipsum, and more. The tool maps TypeScript field names to appropriate Faker methods.

Future idea: Add knobs and sliders to customize data generation (e.g. age ranges, locales).

Try It!

Paste this into the generator:

interface User {
  id: string
  name: string
  age: number
  contact: [phone: string, email: string]
  isActive: boolean
  role: 'admin' | 'user' | 'guest'
  tags: string[]
  children: [Child] | [Child, Child]
  address: Address
}

interface Address {
  street: string
  city: string
  country: string
}

type Child = `${First} ${Digit}`
type First = string
type Digit = '0' | '1' | '2' | '3' | '4' | '5' | '6' | '7' | '8' | '9'

Hit generate and see what happens! ✨

One more thing

I almost forgot, it has a dark mode! 🌙 How awesome is that!

Open Source & Looking for Ideas

The project is open source and I'd love contributions!

Support for more TypeScript patterns
Integration with more faker mocking functions, or other mock data libraries
Custom field mapping configuration
CLI tool
VS Code extension
API

I am happy to receive Suggestions for improvements and Bug reports.

Check it out: https://github.com/ackvf/ts-mock-generator

Wrapping Up

This started as a weekend project and turned into something I just play with all day. It's snappy and so delightful to use. 🤗 If you work with TypeScript and need mock data, give it a shot!

Found it useful? Star the repo, share it, or drop suggestions in the comments! 🙌

Stack: Next.js, TypeScript, Faker.js, Tailwind CSS, Shadcn/ui, Base UI

Tags: #typescript #webdev #testing #opensource #mockdata #json

Don't get scammed on an interview.

Qwerty — Wed, 26 Nov 2025 00:19:00 +0000

Don't get scammed on an interview.

So… I just went through two “interviews” that turned out to be scams. Both followed the same playbook:

they gave me a repo link, told me to share my screen, and asked me to install + run the project during the call.

On a first interview.

That alone feels off, right? But these scammers have gotten smarter. They’ll do the whole friendly intro, talk about the company, show you a project, ask about your past work — the whole thing feels legit until suddenly you’re the one walking them through a repo on your machine.

Luckily, I had a gut feeling something wasn’t right. I isolated everything, sandboxed the repo, and ended up finding a bunch of red flags… including a hidden script or a fake MetaMask popup that looked exactly like the real extension.

The worst thing is that while I already knew this was a scam, I heard another scammer in the background doing the same thing to another poor dev. I even saw their cursor moving in a shared Figma file. However, it was hilarious to watch the panic after I wrote this:

Anyway, this wasn't my first experience either, but seeing how profitable these scams must be, I finally decided to write about it to help other devs avoid getting burned.

Take this as advice from a friend who just dodged a bullet.

tl;dr

If you are interest in knowing how to spot these scams, read on, otherwise skip directly to the How to protect yourself section for a detailed guide, or skip all the way to the step-by-step screenshot guide, or just remember these key points:

If you are in a rush to open unknown repos, use GitHub Codespaces or codesandbox with Copilot or another AI integration to analyze the repo for malicious intent and to run it in a safe environment.

GitHub codespace is a VSCode in the browser at github.dev that gives everyone 60 hr free compute time every month - ideal also for scoped home assignments.

Alternatively, you can self-host full VSCode IDE in a docker with code-server, or just use VSCode Server in a docker and connect remotely with any VSCode. You can also try Dev Containers for free and completely local, that do all that, batteries included.

Btw, what is the difference between github.dev and vscode.dev?

Why listen to me?

I am a software developer with 10 years of experience interviewing at real companies. As a remote contractor, I've worked on 20 projects ranging from few months to couple years and the number of interviews done is significantly higher (20 just this year). Here are the red flags I noticed.

The red flags

Flags hidden in plain sight.

You saw the crimson early.

The tears weren’t yours.

Different ways to get scammed and the First Red Flags

There are multiple ways how these scammers try to invade your system or steal your credentials.

some try to scam you right away via fake conferencing apps
some will let you try a demo project even before the interview and ask you to provide "feedback and suggestions for improvement"
some will ask you to clone and run a repo on your machine with the same agenda either as a homework or during an interview
and others try to build trust first by actually conducting a full interview

🟥 Never use unknown conferencing apps

Insist on using Zoom, Google Meet, Microsoft Teams, Telegram or other mainstream apps.
They are available on desktop, mobile as well as in a browser, so you shouldn’t have any issues joining without installing anything. If it was legitimate, they would agree without any problems, especially if e.g. teams don't work for you, so you ask to use another.

Be wary of links like meet.google.com.join.1bc-23e-h4j.com or zoom.us.joinsession.com and similar. Those are legitimately looking malicious sites, the important bit is the domain name before .com (or other TLD) - that’s the actual site you are visiting.

If you are curious, before you open any links, use incognito browser mode without any sensitive cookies or extensions enabled to avoid leaking your credentials.

🟥 If they want you to install software, it's a scam

Never install unknown software on your machine. I have never experienced a real interview that required installing anything beyond mainstream conferencing apps.

🟥 If they give you access to their full repo, it's a scam

If they ask you to inspect their full repo or Figma design document and give feedback, it’s absolutely a scam. No real company would give away their intellectual property like that.

Though, sometimes you are required to do a home assignment, but that’s different. In that case, you get a limited repo or a coding challenge designed specifically for interviews - and this is what this post is about and how to detect this, because you can absolutely get scammed even during these assignments.

🟥 If the offer sounds too good to be true, it probably is

... too good to be true.

These scammers always have multiple “open positions” they’re trying to fill, regardless of your actual skills or experience. Full time, part time, freelance, junior, senior — they’ll claim to have something for everyone. They will promise you a high salary, or claim to have a position that perfectly matches your skills.

🟥 Do a background check on their profile

I was once contacted on LinkedIn by a 72 year old "senior web3 recruiter" who was previously a certified hypnotherapist of 40 years.

The plot twist?

The real guy, an actual hypnotherapist, died 2 years prior which I found after reverse-image searching his profile picture.

🟥 The second Red Flag: The way They talk to you

If you get past the initial obvious signs mentioned above, there are more subtle red flags to watch out for.

One thing I didn’t expect is how predictable the communication style of these scammers is.

If you pay attention, you can spot them before you ever touch a repo.

Here are the patterns I noticed — and once you see them, you can’t unsee them.

They sound surprisingly unprofessional

You’d expect an experienced recruiter or a senior engineer to speak clearly, guide the discussion, explain the agenda, and set expectations.

With scammers, it’s the opposite:

no real structure
unclear questions
unclear instructions
often poor English and vague phrasing
awkward pauses
jumping between topics without transitions
not really going into depth on anything
proceeding with their script regardless of what you say or fails you make
noisy/bussy room

They’re often inexperienced or too “casual”

Young, impatient, and not used to real hiring processes.

They don’t probe deeply into your experience.

They don’t ask follow-ups.

They don’t explore your reasoning.

It feels like someone trying to pretend they’re running an interview rather than someone who actually does this for a living.

They rush the introduction phase

The intro is usually the weakest part of the whole act, because scammers:

don’t understand the actual product
can’t answer deep questions
aren’t trained in interview structure
don’t want to waste time

They don’t actually care about your career or fit

They don't care about:

how you think
how you solve difficult challenges
how you collaborate
what kind of developer you are
whether you’re someone they want to hire

Real interviewers are curious about you, your story, your skills and thought process.

They have a short fuse

If you hesitate, ask questions, or try to clarify anything, they get impatient or even annoyed. Real interviewers are patient and willing to explain things clearly. You can try it by simply playing dumb on purpose for a bit and see how they react - it's really fun actually :-).

They overpraise you excessively

Another thing I noticed: when I showed my portfolio, the reactions were overly enthusiastic to the point of feeling fake — constant “wow, amazing, man!” every few seconds. Experienced interviewers might acknowledge good work, but they don’t oversell it like that. Excessive praise is just another social engineering tactic to build quick trust and rush you past your instincts.

In a real interview, you should feel like you’re having a professional conversation with someone who knows what they’re doing and is genuinely interested in the details you share.

If it feels off, it probably is.

Disabled camera or filters are a red flag too

If they don’t turn on their camera, that’s a big warning sign. Real interviewers and developers always show their face and expect the same from you at any stage of the interview. I can’t stress this enough: in my ~50+ interviews, I have never had a single call where at least one person didn’t already have their camera on when I joined. While this isn’t a hard rule, it’s a very strong convention in tech interviews.

And if they do turn it on, pay attention to their background, lighting, and overall professionalism. A messy or odd setup can be a sign they’re not who they claim to be. Even recruiters working from home make an effort to look presentable.

Also be wary of AI filters or avatars — they’re terrifyingly good now. Watch for subtle glitches: unnatural blinking, strange eye contact, stiff facial expressions, lips not syncing to speech, or weird distortions when they move and turn their head. If something feels off visually, trust that feeling.

🟥 The last Red Flag: They make You share Your screen

Real companies don’t ask you to share your screen in the first interview - be it to show your portfolio (they usually want portfolio ahead of time, with your CV), run code, or anything else.
Real companies don’t ask you to clone and run their repo in the first 45 minutes of meeting you.

They especially don’t sit back and watch your screen like it’s a Twitch stream.

In any normal code-review-style interview, it's usually a second or third round and the interviewer sets up a shared multi-user coding environment using codesandbox or the likes and only then maybe ask you to share your screen, or just connect to a session. You don’t open unknown software on your machine while strangers watch your desktop icons, extensions, file paths, and environment variables.

If someone insists on you running their repo early in the process: that’s your warning.

How to protect yourself

Even in completely legitimate situations, if you find yourself in a situation where you’re being asked to run unknown code, here are the steps to protect yourself:

You can skip this section if you only want to see the step-by-step guide.

Use isolated coding environment

There are multiple ways to do this, depending on your needs, resources and time.

Dev Containers - VSCode extension

You can explore Dev Containers that come with a VSCode extension, it runs locally in Docker or in cloud, batteries included.

Linux in Docker

Another option is to set up a Docker image yourself with a prebuilt image or clean Linux. Experience very similar to just using WSL for development. You will connect "remotely" from your local VSCode over SSH, which triggers the installation of VSCode Server. From there you can install node and other dependencies, and clone and run the repo; fully isolated.

Self-hosted VSCode in Docker

As we move to the cloud, you can self-host full VSCode IDE in a docker with code-server. This will give you web-based VSCode that you can access from any browser, any device, anywhere. Great for occasional development from a tablet.

GitHub Codespaces or codesandbox

Some might argue that Docker is not 100% secure, so that leaves us with Virtual Machines or 3rd party cloud providers, such as GitHub Codespaces or codesandbox.

GitHub Codespaces provides 60 hours of free compute time every month, which is more than enough for scoped home assignments or interviews. It’s a full VSCode in the browser at github.dev or vscode.dev.

What is the difference between github.dev and vscode.dev?
Difference between the two is that github.dev is more integrated with github: automatic login, and some github related extensions, including Github Copilot and the github visual theme. It can load any github repo from the command prompt, or by pressing . on a keyboard while on any github page, or just by replacing the github.com with github.dev in the URL.
vscode.dev requires you to log in if you want to synchronize your settings and extensions and it can load github and azure repos in a similar way by appending the full url vscode.dev/dev.azure.com/myRepo.

However, in my experience it is much faster to just use github.dev for any repositories even outside of github such as bitbucket or gitlab, because a Codespace requires you to start from a repo on github or to create a new one, and https://github.com/github/dev is such a minimalistic repo that loads with github.dev by default.

Analyze the repo with AI tools

Before you run any unknown code, use AI tools to analyze it for malicious intent. You can use GitHub Copilot, ChatGPT, or other AI assistants to review the codebase quickly. All of them are available as extensions in VSCode, so you can use them directly in your isolated coding environment. GitHub Copilot is also available for free in GitHub Codespaces.

Use Agent mode to scan the whole codebase
- A simple prompt "Analyze this codebase for any malicious scripts, hidden backdoors, or suspicious behavior. Summarize your findings." should do.
Ask specific questions about dependency files, configuration scripts, or parts of the code that seem suspicious.

Tip: Watch for unknown dependencies, suspiciously large files and obfuscated code. The following is a pretty common code injection. A normally looking tailwind.config.js file, but there was a malicious one-line code inserted after at least a thousand spaces, completely unnoticeable to a naked eye.

Step-by-step guide that is completely free and instant

This guide uses GitHub Codespaces. It's free and instant to start. You get a full VSCode IDE in the browser with terminal access, extensions, GitHub Copilot, and everything you need to inspect and run the code safely.

docs: Codespaces: Free quota
docs: Codespaces: Free unit consumption multiplier and pricing
docs: GitHub Copilot Free
account: Billing / Usage dashboard

1. Open the repo link in your browser

If it is a github repo, while browsing it, hit . on the keyboard or replace github.com with github.dev in the URL, which will launch the repo in github.dev browser VSCode.
If not, copy the clone command (URL).

2. Start a Codespace on github.dev

If the TERMINAL button doesn't appear, open Command Palette and find

Codespaces: Continue working in New Codespace

Choose 2 cores for 60 hours of free computing time every month. Double the CPU halves the free hours.

3. Clone the repo in the Codespace terminal (optional)

If it was not a github repo, clone it in the terminal:

git clone <REPO_URL>
cd <REPO_NAME>

4. Analyze the code with GitHub Copilot

Open Command Palette and find

Chat: Open Chat

Switch to Agent mode. and ask it to analyze the codebase for malicious intent:

Analyze this codebase for any malicious scripts, hidden backdoors, or suspicious behavior. Summarize your findings.

If you find something, then that's your cue. You can use deobfuscation utilities to get more sense of the code or ask AI, if you are curious, but you should exit the interview.

beautifier.io - online code beautifier

5. Run the code safely

Follow the project’s README instructions to install dependencies and run the project. Since you’re in an isolated Codespace, any malicious code will be contained within this environment. But still be cautious before you enable access to any extensions.
My suggestion would be to run it in a separate Browser profile or Incognito window without any sensitive cookies or extensions enabled to avoid leaking your credentials, such as Metamask or other web3 wallets.

You will see a link in the terminal to open the running project in a new browser tab if it is a web application.

6. Clean up

When you are done, stop the Codespace to avoid consuming more hours and delete the codespace from your GitHub account as it consumes GB-hours. (You get 15 GB for a month for free.) The codespace gets deleted automatically after 30 days of inactivity, too, if you forget.

Check the usage in the dashboard.

If you have already closed the Codespace tab, you can find it at https://github.com/codespaces.

Final thoughts

I don't want to waste any more of your time. This already blew beyond few simple screenshots into a more detailed guide, but I think it was worth it.

Thank you for reading and please, do share it with your junior developer friends to help them avoid getting scammed. The scammers keep improving and move into organized operations, so it must be profitable for them, and with the rise of AI, they will get even more sophisticated.

Cheers! 🥂😎

Qwerty, signing out.

Veni
Vidi
Abii

Vici

(I came; I saw; I left; I won)

DEV Community: Qwerty

Automatically hide _assets folders in Obsidian (until you need them)

What this snippet does

"Install" (Obsidian CSS snippet)

The CSS snippet

How it works (quickly)

Known drawbacks (and why they happen)

1) If there are multiple _assets folders in the active folder…

2) “Hidden” folders still affect virtualization / scrolling

Recommended usage pattern

What I might build next

Credits / inspiration

Delightful TypeScript Mock Generator - Generate Realistic Test Data from Your Interfaces

Why I Built This

The Cool Parts

1. Smart Field Recognition

2. Type Names Matter (The Secret Sauce!)

3. "Invalid" Patterns That Work Beautifully

4. Reproducible Results

5. Advanced TypeScript Support

6. JSON ↔ TypeScript

Real Usage Examples

Under the Hood

About Faker.js

Try It!

One more thing

Open Source & Looking for Ideas

Wrapping Up

Don't get scammed on an interview.

Don't get scammed on an interview.

tl;dr

Why listen to me?

The red flags

Different ways to get scammed and the First Red Flags

🟥 Never use unknown conferencing apps

🟥 If they want you to install software, it's a scam

🟥 If they give you access to their full repo, it's a scam

🟥 If the offer sounds too good to be true, it probably is

🟥 Do a background check on their profile

🟥 The second Red Flag: The way They talk to you

They sound surprisingly unprofessional

They’re often inexperienced or too “casual”

They rush the introduction phase

They don’t actually care about your career or fit

They have a short fuse

They overpraise you excessively

Disabled camera or filters are a red flag too

🟥 The last Red Flag: They make You share Your screen

How to protect yourself

Use isolated coding environment

Dev Containers - VSCode extension

Linux in Docker

Self-hosted VSCode in Docker

GitHub Codespaces or codesandbox

Analyze the repo with AI tools

Step-by-step guide that is completely free and instant

1. Open the repo link in your browser

2. Start a Codespace on github.dev

3. Clone the repo in the Codespace terminal (optional)

4. Analyze the code with GitHub Copilot

5. Run the code safely

6. Clean up

Final thoughts

1) If there are multiple `_assets` folders in the active folder…