Welcome ! You’ve just joined YOUR_DREAM_COMPANY, a fast-growing technology company. They have a Python Flask application, that powers their services, and like any real-world project, it needs constant care, new features, and a bit of cleanup.
Instead of a traditional tutorial where you just copy-paste code, this workshop puts you in the driver’s seat. You’ll work alongside an AI Agent (Google Antigravity or Gemini CLI) to complete 10 common developer tasks. It’s a hands-on simulation designed to show you how “Agentic workflows” can help you build and maintain software more effectively.
Choose Your Path
There are two ways to navigate this workshop. Pick the tool that fits your workflow:
- Google Antigravity: An Agent first development tool that also provides a visual, browser-based editor. It’s perfect if you want to see your code and the AI’s plan side-by-side.
- Gemini CLI: A powerful terminal tool. This is for the “Terminal Warriors” who prefer using the @ symbol to give the AI specific context directly from the command line.
The 10 Activities
We’ve broken the workshop down into 10 bite-sized activities. Each one represents a task you’d likely face in a typical week (or in today’s fast paced AI world, within a day) at a software job.
01 Vibe-Coding: Build a “Zen Travel Planner” UI with smooth animations.
02 Documentation: Use the agent to explain and document complex logic.
03 Feature-Building: Add a “Low Stock Alert” system to the dashboard.
04 Refactoring: Clean up messy “spaghetti” code into organized modules.
05 Testing: Generate a full test suite to ensure the Finance API is reliable.
06 Migration: Modernize the codebase to use Python 3.12 standards.️
07 Bug-Fixing: Squash logic errors hidden in the Inventory API.
08 Security: Find and patch vulnerabilities like SQL injection and XSS.️
09 Performance: Fix slow code and nested loops to speed up sales reports.
10 Deployment : Package the app for the cloud with Docker and CI/CD.
The Workflow: Reproduce, Prompt, Verify
To get the most out of this workshop, we follow a simple three-step loop for every single task:
- Reproduce: Run the app and see the problem (or the missing feature) for yourself.
- Prompt: Use your AI agent (in the IDE or CLI) to propose and apply a fix.
- Verify: Run the code or the tests to make sure everything works as expected.
By following this cycle, you aren’t just letting the AI write code , you are directing it to solve specific problems.
Ready to start?
The best way to learn is by doing. Everything you need — including the code, the environment setup, and the step-by-step guides — is waiting for you in the repository.
Explore the agy-workshop on GitHub
GitHub – rominirani/agy-workshop: A comprehensive 10-project curriculum for developers to learn agentic coding mindsets. Covers rapid UI generation, documentation, refactoring, security patching, and performance optimization using state-of-the-art AI coding assistants.
Jump into Activity 01, start “vibe-coding” your travel planner, and see how an AI agent changes the way you build!
Check out the detailed README.md file in the project loop
Feedback
This workshop is an initial start and I would love for you to validate a few things, provide some interesting prompts and even exercises. Feel free to collaborate and send across a PR.
Beyond the Basics: 10 Hands-On Activities for the Agentic Developer was originally published in Google Cloud – Community on Medium, where people are continuing the conversation by highlighting and responding to this story.
Source Credit: https://medium.com/google-cloud/beyond-the-basics-10-hands-on-activities-for-the-agentic-developer-8696804ccff0?source=rss—-e52cf94d98af—4
About the Author
