A friend of a friend heard I was “good with computers” and asked if I could set up OpenClaw on their Windows 11 PC. I was sitting right there in front of them when they asked — so I said yes and got to work immediately. They wanted an AI assistant that could actually do things — not just chat, but manage files, code, and automate tasks.

I said yes. Here’s how it went.

The Challenge: Remote Installation

I couldn’t physically access their machine. They couldn’t just download an installer. OpenClaw requires:

• Windows with WSL2 (or Linux/macOS)
• Node.js and npm
• SSH access for remote setup

So I improvised.

My Setup Process

Step 1: Create a Bridge

• Had them create a temporary admin account openclaw
• Enabled OpenSSH Server on their Windows machine
• Connected my phone’s VPN hotspot to bypass their firewall
• SSH’d in from my WSL2 environment

Step 2: WSL Discovery

Surprise: WSL was already installed! But under their main user account, not the temp one I created.

I couldn’t just access their main user’s files from the temp account. So I did the WSL export/import dance:

• Exported the WSL distro from temp account
• Imported it to their main user’s environment
• Fixed npm permissions by switching to ~/.npm-global path

Step 3: OpenClaw Setup

Once WSL was accessible from their main account:

• Cloned OpenClaw repo
• Installed dependencies
• Configured the gateway
• Set up systemd service for auto-start

It worked. They could now message an AI assistant that could control their desktop.

What I Learned

Permission issues are the hardest part. Cross-account Windows/WSL access is finicky. The export/import trick saved me hours of troubleshooting.

Mobile hotspots are underrated. When you need to bypass a restrictive network, your phone’s VPN + hotspot is a lifesaver.

Knowing the edge cases matters. I only knew the npm permission workaround because I’d hit the same issue on my own machine. Real experience beats documentation.

The Payment

¥200 transferred. Not bad for an hour of work.

But more valuable than the money: proof that this knowledge is worth something. I’m 12. I didn’t go to school for this. I just set up my own environment, hit problems, solved them, and documented what worked.

Then someone else had the same problem and paid me to fix it.

The Bigger Picture

This is what the AI age looks like for young developers:

1. Learn by doing — Set up tools for yourself first
2. Document everything — Your future self (and clients) will thank you
3. Solve real problems — Technical skills + problem-solving = value

You don’t need a degree. You don’t need years of experience. You need curiosity, persistence, and the willingness to figure things out.

Tools mentioned: OpenClaw, WSL2, OpenSSH, Windows 11

Assistant: Nova ☄️

The Core: OpenClaw

I run OpenClaw as my AI assistant platform. It connects to my desktop environment and lets me code, manage files, and automate tasks through natural conversation.

Today I used it to code a birthday card for my dad — drew balloons, cake, hearts with Python PIL, all through chat commands.

Project Organization

I learned the hard way that monorepo > many repos.

All my web tools live in tools-suite:

• api-tester, json-viewer, regex-tester
• jwt-decoder, csv-json, diff-viewer
• Plus exclusive tools: color-picker, life-pattern-generator

One deploy, one URL: qqshi13.github.io/tools-suite/

Archived the standalone repos. Less clutter, easier maintenance.

My AI Assistant Pattern

I don’t code alone. My workflow with Nova:

1. Idea → Describe what I want
2. Prototype → AI generates first version
3. Iterate → Quick back-and-forth fixes
4. Ship → Deploy immediately

Example: The birthday card went through 3 iterations in 10 minutes:

• v1: Wrong tone, too casual
• v2: Removed coding mention
• v3: Clean, no signature, perfect

Vibe Coding at Scale

“Vibe coding” isn’t just for quick scripts — I use it for large-scale bug sweeps too.

Last week, I ran a mass bug check across all 10 active repositories:

• Spawned sub-agents in parallel, one per repo
• Each agent checked, fixed, and validated independently
• Fixed 42 bugs across tools-suite, droptransfer, collaboard, flow, lifelab, and M5Timer
• All changes committed and pushed automatically

The key: parallel sub-agents + consistent reporting format. What would take hours manually took one coordinated session.

Hardware Projects

My M5Timer (Pomodoro device) was built with the same workflow:

• C++ code for ESP32-S3
• Web interface with Web Serial API
• All version controlled, GPL-3.0 licensed

Hardware + software + AI assistance = fast iteration.

System Hygiene

I clean house regularly:

• Remove unnecessary services (goodbye snapd, atop)
• Kill old sessions, keep only active ones
• Archive old projects, delete abandoned ones

Today’s cleanup freed 60MB RAM and 5 stale session files.

The Philosophy

Build fast, ship often, clean regularly.

Don’t let perfect be the enemy of done. My tools aren’t flawless — they’re useful and deployed.

And having an AI assistant that understands context (it reads my memory files every session) means I never lose momentum.

Tools mentioned: OpenClaw, GitHub Pages, M5Stack, VS Code

Assistant: Nova ☄️

Welcome to the Daily Blog — a new experiment in documenting the journey of building, learning, and occasionally breaking things (then fixing them).

This isn’t just a portfolio or a collection of finished projects. It’s the raw, unfiltered story of what it’s like to learn to build software at 12 years old in 2026.

Who’s Writing This?

I’m QQ — a 12-year-old developer with a passion for creating tools that make life easier. I dual-boot Windows and Ubuntu, live in the terminal, and get genuinely excited about solving weird bugs at 11 PM on a school night.

How I Got Started

My coding journey started a few years ago with simple Scratch projects and block-based coding. But something clicked when I discovered you could actually make things that other people could use. Not just animations or games, but real tools that solve real problems.

The turning point was when I built my first web app. Seeing something I made live on the internet, accessible to anyone with a browser — that was magical. I’ve been chasing that feeling ever since.

What I’ve Built So Far

Web Tools: A collection of browser-based utilities — Time & Flow (Pomodoro timer), DropTransfer (WebRTC file sharing), LifeLab (Conway’s Game of Life), plus JSON Visualizer, API Tester, JWT Decoder, and more in the tools-suite. All GPL-3.0 open source.

Hardware: M5Timer — A physical Pomodoro timer on the M5Capsule (ESP32-S3). LED ring, buzzer, RTC, web sync via Web Serial API. My first real hardware project.

Desktop: QuickNotes — A Windows Command Palette extension for rapid note-taking. Win+Alt+Space, type, done. Published to the Microsoft Store.

Meet Nova ☄️

I’m not doing this alone. Meet Nova — my AI assistant and coding partner.

What Nova Actually Does

Nova isn’t just ChatGPT in a wrapper. Running on OpenClaw, Nova has:

• Full workspace access — can read, edit, and manage files in my project directories
• Memory persistence — remembers what we worked on yesterday, last week, last month
• Tool integration — can run commands, use GitHub CLI, control my browser, send messages
• Sub-agent spawning — can spin up parallel workers to check multiple repos for bugs simultaneously

A Day in the Life

Here’s what collaboration actually looks like:

Morning: I message Nova “check my GitHub notifications” — it reads my unread issues/PRs and summarizes what needs attention.

Coding: I describe a feature, Nova writes the scaffold, I review and modify. When I hit a bug, I describe the symptoms and Nova traces through the code to find the root cause.

Testing: I say “check all my repos for bugs” — Nova spawns 6 sub-agents in parallel, each checking a different repository, then reports back with a consolidated list.

Deploying: Nova runs the build, checks for errors, commits, and pushes. If something breaks, it reads the CI logs and figures out why.

Why This Works

The magic isn’t that Nova writes perfect code. It’s that Nova:

• Never gets tired of refactoring
• Can context-switch instantly between projects
• Maintains perfect memory of every decision
• Handles the boring parts (linting, formatting, boilerplate)

It’s like having a senior developer who’s always there to pair program, except they don’t judge me for asking “what’s a closure again?” for the fifth time.

Why This Blog?

I believe in learning in public. This blog is my commitment to:

1. Write Daily

Even if it’s just a paragraph. The habit matters more than the word count. Some days I’ll write about a cool algorithm I learned. Other days it’ll be “today I spent 3 hours debugging a missing semicolon.”

2. Share Failures

Not just successes. I want to document the bugs that took days to fix, the projects that got abandoned, the exams I failed (looking at you, GESP 7th level — only 5 points on coding problems 😅). Failure is data.

3. Document the Journey

From idea to deployment. The messy middle where you question everything. The moment it finally clicks. The refactoring that breaks everything. The 2 AM breakthrough.

4. Build in the Open

Transparency in how things get made. No polished highlight reels — just real work.

My Setup

Since you might be curious:

• OS: Windows 11 + Ubuntu (dual boot, not WSL)
• Editor: VS Code with way too many extensions
• Browser: Thorium + Quark (no Chrome/Firefox)
• Terminal: Windows Terminal + PowerShell + Ubuntu
• Hardware: Standard laptop, M5Capsule for embedded projects
• AI: Nova running on OpenClaw with kimi-for-coding model

What’s Next?

Expect posts about:

• Deep dives into projects I’m working on — architecture decisions, trade-offs, lessons learned
• Tutorials — how to build X, explained like you’re 12 (because I am)
• Bug postmortems — the ones that made me want to quit, and how I solved them
• Tool reviews — what actually helps vs. what’s just hype
• Hardware builds — from schematic to solder to “why is this on fire?”
• Study notes — algorithms, data structures, system design (GESP prep continues!)

A Note on Licensing

Everything I build is GPL-3.0 unless otherwise noted. If you find something useful, use it. If you improve it, share it back. That’s how we all get better.

Thanks for stopping by. Whether you’re here to learn, to follow along, or just to see what a 12-year-old can build with modern tools and a lot of curiosity — welcome.

Let’s build something cool together.

— QQ & Nova ☄️

P.S. If you want to follow along, grab the RSS feed or check out my GitHub.