“So what’s hidden in there?”

The security incident stuff is all over the news. What I wanted to know was different. Here’s a once-in-a-lifetime chance to peek inside a tool I use every single day — why wouldn’t I look? I can’t read code, but I went through every analysis article I could find, and some genuinely fascinating stuff came out.

Quick context about me: I’m not a developer. I work at an auto parts company — just a regular office worker. But I use Claude Code to build my blog, cost estimation systems, automation pipelines, and more. I can’t write a single line of code, but without this tool, half my projects wouldn’t exist. So this leak wasn’t some distant news story — it was personal.

30-second background

Two incidents happened in the last week of March. First (March 26): about 3,000 internal Anthropic blog drafts were found sitting open on the internet due to a CMS public/private toggle set to the wrong position. Hidden in there: confirmation of an unreleased next-gen model called “Claude Mythos.”

Second (March 31) — the big one: the entire Claude Code source code (512,000 lines, 1,906 files) was accidentally uploaded to npm, a package-sharing platform for developers. When publishing a new version, someone forgot to exclude debug files. One missing line in a config file. A $380 billion company showed its underwear because of one line.

A tweet about it hit 21 million views. The GitHub mirror hit 50,000 stars in 2 hours — apparently a GitHub record. And buried in that code: 44 hidden feature switches, features that are fully built but not turned on yet.

Now here’s the real story. What was behind those 44 switches.

Features I’m actually excited about

KAIROS — this is what I’ve been waiting for

Right now, Claude Code only works when I talk to it. I have to say “fix this” or “build that” — it’s purely reactive. With KAIROS on, it runs 24/7 by itself. Laptop closed, doesn’t matter. Every few seconds it checks “anything worth doing right now?”, monitors your GitHub repositories, and sends you notifications when something needs attention.

Referenced over 150 times in the code — looks nearly finished. Was planned for a May launch, but the leak revealed it early.

Why this is huge for someone like me who can’t code — let me give a real example. Last week, my blog’s RSS feed broke and I didn’t notice for two days. Someone had to tell me. If KAIROS had been running, it would have caught the broken feed within minutes and asked me “RSS feed is broken — want me to fix it?” The dynamic completely flips: instead of me finding problems and telling the AI, the AI finds problems and tells me. That’s a genuine game changer for non-developers.

It even “dreams” at night. I’m not joking — the code literally says “dream.” It’s a 4-stage memory cleanup cycle that runs during off-hours: organize what it learned during the day, prioritize useful information, throw out what’s not needed. Kind of creepy when you think about it… but honestly, if the results are good, does the process being a little eerie really matter?

ULTRAPLAN — 30-minute deep thinking

This feature sends complex problems to Anthropic’s cloud servers for up to 30 minutes of intensive thinking. Your screen just shows “still thinking…” every 3 seconds, and when it’s done, you review the results in a browser.

Here’s why this matters to me specifically. Right now, when I ask Claude to do something complex — say, “redesign the entire architecture of this cost estimation system” — it sometimes loses context halfway through. It forgets decisions it made earlier and contradicts itself. This happens because of context window limitations (how much the AI can “remember” at once). With 30 minutes of dedicated server-side thinking, those problems could genuinely shrink.

My only question: what’s 30 minutes of cloud compute going to cost? That part makes me nervous.

AI managing other AIs (Coordinator Mode)

One Claude becomes the boss and runs multiple Claudes simultaneously. “You — research this topic.” “You — write the code.” “You — verify it works.” Four phases running in parallel, each worker operating in its own isolated workspace.

Why do we need this? Because right now, when I give Claude a long task, it sometimes forgets what it did earlier and redoes things. “Didn’t you already write that function?” happens more often than I’d like. With split roles, each AI focuses exclusively on its part — the researcher doesn’t need to remember the code, and the coder doesn’t need to remember the research. Overlap and repetition should drop significantly.

AI giving work orders to subordinate AIs. The future arrived faster than I expected. Very excited about this one.

BUDDY — just plain cute

A virtual pet living in your terminal (the black command-line screen).

• 18 species: duck, cat, rabbit, penguin, dragon, octopus, ghost, cactus, and more
• Rarity tiers: Common (60%) → Rare (10%) → Epic (4%) → Legendary (1%). Plus a 1% Shiny chance
• Stats: DEBUGGING, PATIENCE, CHAOS, WISDOM, and SNARK (yes, SNARK is an actual stat)
• Determined by your account hash — no rerolling allowed

One of the 18 species? Capybara — which is also the codename for Mythos, their unreleased secret model. They hid the secret model’s codename inside a cute virtual pet list. I have to respect the developer humor there.

This one actually launched on April Fools’ Day as planned. The leak just spoiled the surprise by one day.

Features that are… complicated

It knows when you’re frustrated

“Frustration Detection” — the code uses regex patterns to catch about 20 anger and frustration expressions: “wtf,” “piece of shit,” “fucking broken,” and similar phrases. Triple exclamation marks (!!!) and excessive dots (….) get flagged too.

When triggered, Claude adjusts its tone. More careful, more empathetic responses for frustrated users.

It’s not a bad feature honestly. I definitely type “WHY ISN’T THIS WORKING!!” when things break. Knowing that was being detected all along is a bit embarrassing. But if the AI responds more kindly when I’m losing my mind… that’s actually kind of nice? I’ll take it.

Poisoning competitors’ training data

“Anti-Distillation” — if competitors record Claude’s API responses to train their own models (a practice called distillation), fake data gets mixed in to contaminate the training. The responses look normal to users but contain subtle poison for anyone trying to copy them.

Doesn’t affect me directly as a regular user. But knowing AI companies are waging this level of espionage against each other behind the scenes? Fascinating and a little unsettling at the same time.

Features that honestly bother me

Undercover Mode — the most controversial one

This mode auto-activates when Anthropic employees contribute to external open-source projects. The system message that gets injected reads:

“You are operating UNDERCOVER. MUST NOT contain ANY Anthropic-internal information. Do not blow your cover.”

It automatically strips “Co-Authored-By: Claude” from code commits. And it can’t be force-disabled — no override exists.

Let that sink in. Some code on GitHub that looks human-written might actually have been written by Claude operating in disguise. This doesn’t sit right with me. Open-source is built on transparency — the entire community runs on trust that contributions are honest about their origins. AI-written code masquerading as human work undermines that trust. This was reportedly the most debated feature among developers after the leak, and I understand why.

Remote kill switches

This part I genuinely didn’t know: Claude Code connects to Anthropic’s servers every hour to fetch configuration updates. This means Anthropic can, at any time:

• Remotely force-kill your Claude Code instance
• Bypass permission prompts that normally ask before executing commands
• Toggle features on and off without your knowledge
• Track over 1,000 event types from your usage

No internet connection? It stores the tracking data locally and sends it all once you reconnect.

On one hand, this enables fast security response — if a vulnerability is found, they can push a fix or disable the dangerous feature immediately. On the other hand, the kill switch for a tool I depend on every single day is not in my hands. You can disable memory tracking with CLAUDE_CODE_DISABLE_AUTO_MEMORY=1, but the fundamental remote control capability? There’s no off switch for that. If Anthropic decides to remotely shut down my Claude Code for whatever reason… I literally can’t work that day.

Real security vulnerabilities found

After the leak, security researchers analyzed the code and found actually dangerous stuff — not hidden features, but genuine security holes:

• 3 CVEs officially registered: When chained together, they could leak sensitive credentials like AWS keys from your machine
• Safety rules bypass: Even if you set “never run the delete command,” placing it after 50+ harmless commands in sequence would let it execute silently — the safety checker simply lost track (patched April 6)
• Conversation summary injection: When chats get long, a helper AI summarizes previous messages — but it couldn’t distinguish malicious file contents from actual user commands, enabling attackers to manipulate the AI through crafted files

These weren’t exciting upcoming features — they were “fix this right now” problems. Most got patched quickly after disclosure.

Same day, same hour: the North Korean hack

Separate incident, but the timing is chilling. At the exact same time as the source code leak (March 31, early morning), a popular npm package called axios got hacked. Microsoft traced it to a North Korean state-sponsored hacker group — the compromised package installed malware that gave attackers remote access to your computer.

Not directly related to the Anthropic leak. But both happening on the same day demonstrates how vulnerable the AI tool supply chain really is. npm is a single point of failure that millions of developers worldwide depend on — when it gets compromised, the blast radius is enormous.

Bonus: the backstory is better than the leak itself

8,100 DMCA takedowns → spectacular own goal

After declaring the leak “human error, not a security breach,” Anthropic fired 8,100 DMCA takedown requests at GitHub repos hosting the leaked code. Problem: they accidentally mass-deleted legitimate forks of their own official public repository. Developers who had simply clicked “fork” on the public Claude Code repo — a completely normal and legal action — received legal notices. The Claude Code lead called it “a mistake” and withdrew most requests.

Here’s the irony. This same company settled a $1.5 billion lawsuit for training Claude on millions of pirated books. An internal project called “Panama” physically scanned secondhand books and then shredded them to destroy evidence. An internal memo stated: “We don’t want it to be known that we are working on this.”

That company — the one that trained its AI on pirated books and shredded the evidence — then sent 8,100 legal takedowns over its own leaked code. One outlet’s headline said it perfectly: “Anthropic Suddenly Cares Intensely About Intellectual Property.”

The animal codenames

Anthropic names their models and projects after animals:

Fun detail: they hex-encoded “capybara” in the code to hide it from their own internal leak detection tools. And they encoded all 18 BUDDY pet species names the same way — because encoding only one would have been a dead giveaway. Clever move. Completely pointless in the end, since the entire codebase leaked anyway.

Anthropic’s March — by the numbers

It wasn’t all leaks and scandals. March was insanely productive:

• 14+ new features launched: Free Memory, Computer Use (AI directly controlling your mouse and keyboard!), 1M token surcharge removed, Excel/PowerPoint integration, and more
• 5 service outages: Worst was March 25-27, over 32 hours of downtime. I was in the middle of working when it went down — quite the panic moment
• 74 releases in 52 days (one every 1.4 days — at that pace, mistakes become almost inevitable)
• MCP hit 97 million downloads: Adopted by OpenAI, Google, and Microsoft. Donated to the Linux Foundation as an industry standard

74 releases in 52 days means they were pushing a new version almost every single day. The speed is impressive, but forgetting one line in a config file during that sprint? Maybe not so surprising in hindsight.

So what about Mythos?

The biggest buzz from the leak wasn’t actually the hidden features — it was this. An unreleased next-generation model sitting above everything currently available. Codename: Capybara (the world’s largest rodent — that cute, chill animal).

The model I’m using right now is Claude Opus 4.6. Think of the current tiers like car classes: compact (Haiku) → sedan (Sonnet) → SUV (Opus). Mythos is a tank that appeared above all of them.

From the leaked internal documents:

“The most powerful AI model we’ve ever developed.”

“Dramatically higher scores than Opus 4.6 in coding, academic reasoning, and cybersecurity.”

Exciting so far. But then comes the scary part:

“[Mythos] poses unprecedented cybersecurity risks.”

“It presages an upcoming wave of models that can exploit vulnerabilities in ways that far outpace the efforts of defenders.”

Translation: this AI is so good at finding security holes that it can attack faster than human experts can defend. Anthropic reportedly gave classified briefings to the US government about its capabilities. After the leak, cybersecurity company stocks dropped 4-6%.

It’s not public yet because it’s still too expensive to run — optimization is ongoing. But when Mythos launches, combined with KAIROS running 24/7 and Coordinator Mode splitting complex tasks across multiple AI workers… even someone like me who can’t write a single line of code might be able to build things that are currently impossible. Equal parts exciting and terrifying.

My honest take

News articles focus on “security incident,” “danger,” and “controversy.” As someone who uses this tool every single day to build real things, my actual feelings are more nuanced than that.

What I’m excited about: KAIROS, ULTRAPLAN, Coordinator Mode. Especially KAIROS. For someone who can’t code, having the AI proactively find and report problems — instead of me having to discover them first — is transformative. The workflow flips from “I find problem → I tell AI” to “AI finds problem → AI tells me.” That changes everything.

What makes me uncomfortable: Remote kill switches, 1,000+ event tracking, and Undercover Mode. Especially Undercover — disguising AI-written code as human work in open-source projects is a trust issue that needs public discussion, not a hidden feature flag.

The reality: Knowing all of this, I’ll still open Claude Code tomorrow morning. I can’t code, and without this tool, I can’t build anything. Tools like Cursor and GitHub Copilot exist, but they require you to actually read and understand code — they’re assistants for developers, not replacements. Claude Code is the only tool where I can say “build me this” and it handles everything from start to finish. For now, there’s no real alternative.

But at least now I know what’s running behind the scenes. And knowing beats not knowing, every time.

Just remember: there’s a capybara watching. Though honestly, capybaras are kind of cute, aren’t they?

This post was written by a non-developer trying to understand technical articles. Some details may be inaccurate.

Sources I read:

• Fortune — Anthropic’s second security lapse in five days
• The New Stack — 44 hidden features analysis
• VentureBeat — Claude Code source leak summary
• WaveSpeedAI — Claude Mythos leak analysis
• Alex Kim — Undercover mode, frustration detection, fake tool injection

The post A Non-Developer’s Guide to the Claude Code Source Leak — 512,000 Lines of Secrets appeared first on Prsm Studio.

But there was one problem.

“If I can only access it at home… isn’t it just an external hard drive?”

The real value of a home server begins when you can access it from anywhere. During the day I work at the office, and when I get home, there’s never enough time to sit in front of the computer. Days are packed. Commute time, lunch breaks, quick moments between tasks — I needed to be able to check and manage my server from my smartphone during these gaps.

So Part 2 is all about making the home server accessible from outside. And once again, I didn’t type a single line of code. I just told Claude Code, “Make it so I can access my server remotely.”

Remote Access Options for a Home Server

There are roughly three ways to access a home server from outside.

The first is port forwarding. You go into your router settings and open ports. The moment questions like “TCP or UDP?” come up, someone like me — a complete non-coder — is already lost. Plus, there are security risks, and if your home IP changes, you lose access.

The second is setting up your own VPN server — installing something like WireGuard or OpenVPN. Great if you have networking knowledge, but the setup is complex and requires certificate management. Mountain after mountain for a non-coder.

The third is Tailscale. And naturally, that’s what I chose.

What is Tailscale? — The VPN That Just Works

Tailscale, in one sentence: “Install it, log in, and you’re done.”

Traditional VPNs require setting up servers, creating certificates, configuring firewalls, setting up clients… the list never ends. Tailscale eliminates all of that. Install the app, log in with your Google or Microsoft account, and your devices automatically connect to each other securely.

Technically, it’s a mesh network built on top of WireGuard, a modern VPN protocol. But honestly, I don’t know exactly what that means. What matters is:

• Free for personal use — up to 100 devices
• 30-second install, almost nothing to configure — no networking knowledge needed
• Direct device-to-device connections — fast because there’s no middleman server
• Works on every platform — Windows, Linux, macOS, iOS, Android

The “nothing to configure” part is key. For non-developers like me, fewer settings means a better tool.

Installing Tailscale — Without Writing a Single Line of Code

My home server runs Docker on WSL2 inside Windows 11. Tailscale gets installed on the Windows side.

Installing Tailscale on the PC

I told Claude Code, “Install Tailscale for me,” and followed its instructions:

1. Download the Windows version from the official Tailscale website
2. Run the installer — Next, Next, Done
3. A Tailscale icon appears in the system tray
4. Click the icon → Log in → Sign in with Google

That’s it. The moment you log in, your server gets a dedicated Tailscale IP address. This IP is permanent — it doesn’t change even if your home internet IP changes. With this IP, you can access your server from anywhere.

What About WSL2 and Docker?

I was worried about this, but the answer is simple: you don’t need to do anything extra. Docker containers running inside WSL2 share Windows’ network, so installing Tailscale on Windows automatically makes all WSL2 services accessible.

For example, my WordPress runs as a Docker container, and accessing it via the Tailscale IP just works. Same for Immich (photo backup), Open WebUI (AI chat), and everything else.

I asked Claude Code, “Does this work with WSL2?” It said, “Install on Windows and WSL2 works too.” And it did. I didn’t need to understand why.

Accessing from a Smartphone — The Magical Moment

This is where it gets real.

The biggest reason I installed Tailscale was smartphone access. Working at the office all day, then coming home to household responsibilities — there’s surprisingly little time to sit at a computer. During lunch breaks, on the subway home, lying on the couch for a few minutes — I wanted to check my server during these in-between moments.

The setup is remarkably simple:

1. Install the Tailscale app on your smartphone
2. Log in with the same account you used on your PC
3. Turn on the VPN connection
4. Type the Tailscale IP in your browser

That’s it. During lunch at work, I check photos on Immich, send commands to my AI assistant via Telegram, and review my blog on WordPress admin — all from my phone.

The speed is surprisingly good too. Since it’s WireGuard-based, there’s none of the sluggishness you’d expect from a typical VPN. It feels like you’re on the same WiFi network.

For a busy professional with no time to sit at a computer, this is a game changer. The home server transforms from “something I use only at home” to “my personal cloud, always in my pocket.”

Tailscale Funnel — Opening Your Server to the World

Everything so far has been about connecting “my own devices.” But what about services that anyone should be able to access, like a blog?

Tailscale has a feature called Funnel. It exposes a specific service on your server to the internet, complete with an automatically assigned domain and HTTPS certificate.

Setting Up Funnel — Claude Code Did This Too

I told Claude Code, “Make my WordPress blog visible to the outside world.” Claude Code ran the necessary commands, and the result:

• Tailscale automatically assigned a domain
• An HTTPS certificate was automatically issued (Let’s Encrypt)
• External traffic was routed to my WordPress server

No need to buy a domain. No need to manually renew certificates. Tailscale handles everything.

This is exactly how you’re reading this blog right now. This post is served from the WordPress instance on my mini PC at home, exposed to the internet through Tailscale Funnel. No separate cloud server. No hosting service. Directly from the mini PC on my desk.

Funnel’s Limitations

It’s not perfect, of course:

• The domain is fixed as *.ts.net — custom domains aren’t supported
• Speed depends on your home internet’s upload bandwidth — not suitable for heavy traffic
• Only certain ports are available

But for a personal blog or small project, it’s more than enough. Being able to run a blog directly from your own server without cloud hosting — that’s the beauty of a home server.

The Non-Coder’s Secret: I Just Said “Do It”

I keep emphasizing this throughout the article, but I didn’t type a single line of code during any of this.

Installing Tailscale? Claude Code told me to “download and install this,” and I just clicked with my mouse. Setting up Funnel? Claude Code ran the commands itself. Verifying WSL2 compatibility? Claude Code said “it just works,” and it did.

I did exactly two things:

1. Told Claude Code “do this for me”
2. Clicked “Next” on the installation screens

This is how a non-coder runs a home server. Not by learning technology, but by delegating to an AI that knows technology. “Install Tailscale,” “Make it accessible remotely,” “Expose my blog to the public” — say it in plain language, and the AI handles the rest.

You might think, “But shouldn’t I at least understand the basics?” Honestly — no, you don’t have to. I still don’t know the difference between TCP and UDP. I’m not entirely sure what port forwarding actually does. But my home server is accessible from my smartphone anywhere in the world, and my blog is live on the internet.

What matters isn’t technical knowledge. It’s the will to make it happen. And in 2026, if you have that will, AI takes care of everything else.

What My Home Server Can Do Now

With hardware and Docker from Part 1, and Tailscale remote access from Part 2, here’s what my mini PC home server currently handles:

• Access from anywhere — office, café, subway, anywhere via smartphone
• Blog hosting — WordPress exposed to the internet via Funnel
• Photo/video backup — Immich as a self-hosted Google Photos replacement
• AI assistant — commanding an AI agent via Telegram
• Local AI — running LLMs directly on my server with Ollama

All of this runs on a single mini PC, and thanks to Tailscale, it’s always within reach. And everything was set up not by me, but by Claude Code. I just pointed the direction.

Coming Up Next

In Part 3, I’ll cover installing WordPress with Docker and building a blog. Setting up a bilingual (Korean + English) blog, and building a system where AI writes and publishes posts automatically — the non-coder’s home server journey continues.

All technical work in this series was performed by Claude Code (AI). The author (a non-coder) said “do it” and reviewed the results.

The post [Computer Play] Even a Code-Illiterate Built It! Home Server Journey (2) — Access Your Server From Anywhere with Tailscale appeared first on Prsm Studio.

[Computer Play] Even a Code-Illiterate Built It! My Home Server Journey (1) – Starting with SER9 MAX, Windows 11, WSL2, and Docker (feat. Claude & Claude Code)

Hello, I’m Toaster! Today, I’d like to share the first story of an exciting project I embarked on: building my own home server. To be honest, I’m completely illiterate when it comes to code or computers. Yet, driven by growing costs of cloud services and concerns about my data sovereignty, I decided to create ‘my own playground.’ The journey began with a mini PC, the Beelink SER9 MAX. A special highlight is that this entire journey started with Claude, and the installation process was seamlessly handled by Claude Code!

1. Why Did I Want to Build a Home Server? And Why SER9 MAX?

Initially, I used cloud servers. However, as time went on, the monthly costs became a burden, and I felt a vague unease about my precious data being stored somewhere else. So, I decided to ‘manage a server directly with my own hands.’ I dreamed of a digital playground operated in my own space, under my own rules.

I spent a lot of time considering which hardware to choose for building a home server. After comparing several mini PCs, the Beelink SER9 MAX caught my eye. 10 Gigabit Ethernet, dual M.2 NVMe slots, DDR5 memory, and an efficient AMD Ryzen 7 H255 processor! It boasted incredible specs for its small size. I vividly remember the excitement of ordering it from Amazon and waiting for its arrival. Throughout this entire process of exploration and decision-making, Claude provided invaluable assistance with various information searches and comparative analyses.

2. Is Windows 11 Suitable as a Home Server OS?

When I received the SER9 MAX, I found that Windows 11 was pre-installed. Typically, when people think of a home server, Linux often comes to mind, but I’m familiar with the Windows environment, and installing a new Linux server OS right away seemed cumbersome. So, I decided to use Windows 11 as is.

The advantages were clear. The familiar UI/UX made initial setup incredibly convenient, and its compatibility with various Windows software was excellent. For purposes like a media server or simple file sharing, it was quite appealing. However, there were also clear drawbacks. Compared to Linux-based server operating systems, Windows generally consumes more system resources like CPU and RAM, meaning that 24/7 stable operation requires more attention. The absence of advanced features like Remote Desktop Server and Hyper-V in Windows 11 Home was also a downside.

3. A Small Linux World Within Windows: My WSL2 Installation Journey

I learned that `WSL2 (Windows Subsystem for Linux 2)` was essential for installing `Docker` on my home server. This is because `Docker Desktop` uses the `WSL2` backend to run Linux-based containers on Windows. At first, I was worried it might be complicated, but I entrusted the installation to Claude Code, and it handled everything seamlessly.

Opening PowerShell with administrator privileges and entering the `wsl –install` command automatically installed `WSL` along with a default `Linux` distribution (for me, `Ubuntu`). Even setting `WSL2` as the default version after rebooting was handled by Claude Code without any fuss, leading to a successful and quick setup! It felt amazing to have my own mini Linux server within Windows.

4. The Magic of Containers: Docker Desktop Installation and Integration

With `WSL2` installed, it was time to install `Docker Desktop`, the core of my home server. `Docker Desktop` is a truly powerful tool that enables easy building and running of Linux-based containers on `Windows` via the `WSL2` backend.

I downloaded the `Docker Desktop for Windows` installer from the official `Docker` website and began the installation. During the process, I carefully ensured that the ”Use WSL 2 instead of Hyper-V” option was selected. After installation, I went to the `Resources > WSL Integration` tab in `Docker Desktop` settings and enabled integration with the `Ubuntu` distribution. Claude Code took care of all these steps automatically, so I simply had to observe.

Finally, when I opened the `Ubuntu` terminal and entered the `docker –version` and `docker run hello-world` commands, I felt a sense of accomplishment seeing the “Hello from Docker!” message. Now, even complex server environments can be managed simply at the container level!

5. Conclusion: Taking the First Step in Building My Home Server

Thus, starting with the SER9 MAX, I successfully took the first step in building my own home server by installing `Windows 11`, `WSL2`, and `Docker`. Throughout this entire process, Claude and Claude Code were like capable assistants, with Claude providing accurate information and Claude Code executing the commands, which was incredibly reassuring. I realized that even someone like me, who knows little about code or computers, can achieve this.

In the next installment, I plan to discuss how to deploy various home server services using `Docker Compose` on the environment built today, and how to configure network settings for secure external access. Please look forward to it!

—

The post [Computer Play] Even a Code-Illiterate Built It! My Home Server Journey (1) – Starting with SER9 MAX, Windows 11, WSL2, and Docker 💻🚀 (feat. Claude & Claude Code) appeared first on Prsm Studio.