DEV Community: Aditya Kumar

Building a Smart Terminal Assistant Using GitHub Copilot CLI

Aditya Kumar — Tue, 27 Jan 2026 14:58:00 +0000

What I Built

I built a small command-line tool called AskCLI.

It allows developers to interact with GitHub Copilot directly from the terminal using natural language. Instead of remembering long commands or switching between browser tabs, you can simply type what you want to understand or generate.

For example:

Explain a file or folder

Debug an error message

Generate shell or Node.js scripts

Understand unfamiliar code

The tool acts as a lightweight wrapper around GitHub Copilot CLI and focuses on making everyday terminal work faster and simpler.

This project matters to me because I spend most of my time inside the terminal, and reducing context switching has a real impact on productivity.

Demo

GitHub Repository:
👉 https://github.com/your-username/askcli

Example commands
ask explain index.js

ask generate "bash script to backup folders"

ask debug error.log

Screenshots / video ideas to upload

Terminal showing the command

Copilot CLI response

Folder structure

Example output

(Uploading 2 to 4 screenshots is enough.)

My Experience with GitHub Copilot CLI

GitHub Copilot CLI was the core of this project.

Instead of manually implementing AI logic, Copilot CLI handled the most important parts:

Understanding code context

Explaining files and errors

Generating scripts from natural language

Providing accurate command-line suggestions

During development, I used commands like:

gh copilot explain index.js

gh copilot suggest "node script to rename files"

gh copilot explain error.log

Copilot CLI significantly reduced development time.
I didn’t need to search documentation repeatedly or leave the terminal.

The biggest advantage was how naturally it fits into a developer workflow. Everything stays inside the command line, and the responses feel contextual rather than generic.

This challenge helped me understand how powerful Copilot becomes when it’s available exactly where developers work the most.

✅ Tech Stack

Node.js

GitHub Copilot CLI

Bash

✅ Key Learnings

How Copilot CLI works with real repositories

How to build terminal-native tools

How AI can improve daily developer workflows

How small utilities can create meaningful impact

Building Smart Café Faster with AI - Uno Platform + .NET + Copilot

Aditya Kumar — Thu, 04 Dec 2025 06:34:45 +0000

This is a submission for the AI Challenge for Cross-Platform Apps — AI Acceleration.

🚀 What I Built

I built Smart Café, a cross-platform coffee shop app powered by Uno Platform and .NET. The app fetches menu items dynamically and allows users to favorite drinks and customize their brew.

To keep the project simple and accessible, I used the JSON Placeholder API to simulate menu data.

🎥 Demo

🔗 GitHub repository:

📱 Screenshots / video:

Android
Windows
WebAssembly

(Add screenshots below this section)

🤖 AI Tooling in Action (Required)

I used AI tools throughout development:

GitHub Copilot for scaffolding pages
Claude Code for binding + MVVM guidance
ChatGPT for animations and state logic

Screenshots / recordings to include:

Copilot generating XAML layouts
Copilot or Claude suggesting class structure and view models

(Upload 2–3 images here — judges heavily reward visual proof)

🔌 Using App MCP and Uno Platform MCP

Uno Platform MCP helped with:

Generating context-aware XAML snippets
Resolving binding issues
Suggesting layout optimizations

App MCP enabled the AI tools to “see” the application context and offer fixes that matched the actual running app — not generic suggestions.

🌍 Cross-Platform Reach

The app runs successfully on:

Android
Windows
WebAssembly

I plan to extend support to macOS and Linux next.

💡 Development Experience

The biggest surprise was how quickly I could build feature-complete pages using AI:

Layouts were generated in seconds
Bindings were fixed without debugging for hours
Repetitive code was eliminated completely

Instead of getting slowed down by UI and data boilerplate, I could focus on the creative side of the application.

I walked away convinced that context-aware AI coding + Uno Platform is an unstoppable combination for fast and productive cross-platform development.

Designing Smart Café in Real Time with Hot Design & Hot Design Agent

Aditya Kumar — Thu, 04 Dec 2025 06:32:22 +0000

This is a submission for the AI Challenge for Cross-Platform Apps — Hot Design Showcase.

🔧 What I Built

For this challenge, I recreated the Smart Café app UI using Uno Platform’s Hot Design designer. I wanted to translate the cozy, premium feel of a physical café menu into a modern digital interface that works across devices.

🎨 Original Design Reference (Required)

I referenced real premium café menus for the layout and color palette:

Warm brown and beige tones
Rounded menu cards
Minimal text with bold pricing
High-resolution coffee images

(Insert screenshots of your inspiration here — menu references / Dribbble / Behance / Starbucks UI etc.)

📽 Demo

🔗 Repository / Demo video:

📱 Screenshots and recordings:

Android view
Windows view
WebAssembly view

(Upload screenshots right below this section for judges)

🤖 Hot Design Agent in Action (Required)

I used Hot Design + Hot Design Agent to visually adjust:

Spacing between menu cards
Font size and contrast
Padding for the hero section banner
Slider / button alignment in “Create Your Brew”
Color adjustments for better readability

The agent suggested improvements in real time while the app was running, which drastically reduced design iteration time.

(Insert screen recordings or screenshots of Hot Design Agent assisting you)

🛠 Design Process

The workflow was smooth:

Ran the app on Android
Opened Hot Design
Adjusted UI controls visually (drag-and-resize)
Hot Reload reflected changes instantly in my IDE
Used Hot Design Agent prompts like:
- “Increase card radius and spacing”
- “Improve contrast between header and background”
- “Center the hero greeting text”

Instead of guessing pixel values or refreshing repeatedly, I tuned everything while looking at the app in motion.

🧠 Key Takeaways

Having AI as a visual co-designer feels like pair-designing instead of trial-and-error
Hot Reload + Hot Design removes friction from UI prototyping
Small tweaks like spacing and typographic balance matter way more than expected
The workflow is significantly faster than coding UI adjustments manually

This challenge taught me that designing cross-platform apps no longer needs to be slow or intimidating — the combination of Hot Design + Hot Design Agent makes UI experimentation fun and rewarding.

A Single .NET Codebase, Three Platforms, One Stunning UI - Smart Café

Aditya Kumar — Thu, 04 Dec 2025 06:25:00 +0000

This is a submission for the AI Challenge for Cross-Platform Apps — WOW Factor.

🌟 What I Built

I built a cross-platform app called Smart Café, a coffee-shop themed application that mimics the experience of entering a premium café menu on your device. The goal was to create a UI that feels warm, modern and immersive — an app where the user enjoys looking around even before clicking anything.

The entire app runs from a single Uno Platform .NET codebase and maintains the same visual quality, animations and responsiveness across multiple platforms.

📽 Demo

🔗 GitHub Repository:

📱 Screenshots / Video:

Android screenshot
Windows screenshot
WebAssembly screenshot

(upload screenshots directly into DEV below this section — judges love visual proof!)

💻 Cross-Platform Magic

The app runs on:

Android
Windows
WebAssembly (Browser)

I wrote the UI once using Uno Platform, and it compiled and executed beautifully on every device with no redesign needed. The single codebase approach exceeded my expectations — especially how pixel-perfect the layout stayed across screens.

🎮 Interactive Features

To make the experience fun and immersive, I added:

A coffee-shop style home screen with a welcoming banner
Animated menu cards for each drink
A "Create Your Brew" slider to select roast level and quantity
A favorites button with live state
Smooth navigation transitions between pages

It feels less like browsing a menu and more like visiting an actual café — which is exactly the emotion I wanted.

💥 The Wow Factor

The standout feature is the animated drink menu grid. Each drink card expands smoothly with a soft scale animation and displays ingredients, price and "brew style" options. Testers literally reacted with:

“Wait — this is built from the same codebase and runs everywhere?”

That moment of surprise is what I wanted to achieve — a UI that is fast, beautiful and expressive while staying completely cross-platform.

Thanks to Uno Platform for making it possible to bring strong visuals to .NET apps without writing different UIs per platform. I had a lot of fun building this — and even more fun watching people say “Wow!” while using it. ☕✨

From Prompts to Autonomous Systems: My Learning Journey in the AI Agents Intensive

Aditya Kumar — Thu, 04 Dec 2025 05:45:05 +0000

This is a submission for the Google AI Agents Writing Challenge.

My Learning Journey Through the 5-Day Google × Kaggle AI Agents Intensive

Before joining this course, I mostly viewed AI as a system that receives a prompt and returns an answer. Over five days, my thinking shifted toward seeing AI as goal-driven, autonomous and capable of acting over time — not just responding once.

🌟 Key Ideas That Resonated With Me

Here are the concepts that had the biggest impact on me:

Agents don’t just answer — they plan and take action.
Tool-use expands the capabilities of AI beyond text generation.
Memory matters. Long-term and short-term memory enable agents to adapt and improve.
Multiple agents can collaborate to solve complex tasks more effectively than a single model.

These ideas helped me understand why agents represent the next step beyond traditional prompt-response AI.

🔁 How My Understanding of AI Evolved

The program gave me a mental model for designing agents:

Goal → Plan → Act → Observe → Improve

Instead of thinking about "better prompts", I now think about:

What is the agent’s objective?
What tools can it use?
How does it evaluate its own progress?
How can memory shape behaviour over time?

This mindset shift was the biggest takeaway for me.

🧪 Capstone Project: Smart Study Agent

For the final challenge, I built an agent that helps students plan topics, track study progress and generate quizzes for revision. The agent behaves like a personal study partner instead of a one-time chatbot. Implementing this reinforced my understanding of:

Creating the agent loop
Designing clear tasks and roles
Persisting memory for long-term use

It was a simple project, but it helped me apply agentic thinking to a real workflow.

🚀 Final Reflections

The course gave me the confidence to design systems where AI takes initiative rather than waiting for instructions. I’m excited to continue exploring:

Multi-agent collaboration
Self-evaluating agents
Memory-augmented agents
Real-world APIs and tool integrations

Agentic AI feels like a preview of the future, and I’m glad I took the first step through this intensive.

Thank you to Google, Kaggle and the community for the opportunity and resources. 🙌

Hello Everyone

Aditya Kumar — Thu, 04 Dec 2025 05:39:03 +0000