Welcome to the Antigravity workshop! This repository contains 10 activities, all built using Python and Flask, designed to showcase the power of agentic development.

You will explore 10 real-world software engineering tasks across two powerful platforms (Antigravity and Gemini CLI).

Before starting any activity, ensure you have Python 3.8+ installed. Follow these steps for EACH project folder:

1. Enter the activity directory:

   cd 0X-[activity-name]

2. Set up your virtual environment:

   python3 -m venv venv && source venv/bin/activate
   pip install -r requirements.txt

You can complete this workshop using either the Antigravity or the Gemini CLI.

Best for: Visualizing implementation plans, side-by-side diffing, and multi-file orchestrations.

1. Open Workspace: Go to File > Open Folder... and select the agy-workshop directory.
2. Launch Agent Manager: Open the Agent Manager tab (usually at the side or bottom).
3. Prompt: Type your instructions directly into the chat box. Antigravity will create implementation plans and artifacts for you to approve.

Best for: Fast, keyboard-centric iterations and staying in your development shell.

1. Launch Gemini: From your terminal (inside a project folder), simply type gemini.
2. Interactive Mode: You are now in an interactive session with the AI agent.
3. Context is Key: Use the @ symbol to load the files you want to work on.
   • Example: @app.py fix the security vulnerabilities in the search route.
4. Shell Integration: Use the ! prefix to run terminal commands (like pytest or flask run) without leaving the Gemini session.

1. Reproduce: Run python3 app.py and visit http://127.0.0.1:5000. You'll see a completely empty, unstyled page.
2. Prompt: Paste the "Unified Prompt" above into Antigravity or Gemini CLI.
3. Verify: Refresh the browser to see a fully reactive, high-end travel dashboard generated from scratch.

| 02 | Documentation | Knowledge | "@scheduler.py analyze this async logic and generate Google-style docstrings and a comprehensive README.md." |

1. Reproduce: Open scheduler.py. The logic is complex and lacks any comments or documentation.
2. Prompt: Use the prompt above to generate clear, structured documentation.
3. Verify: Check the file for newly added docstrings and look for the new README.md in the 02-documentation folder.

| 03 | Feature-Building | Expansion | "@app.py implement a 'Low Stock Alert' system (< 5 units) with a new dashboard page and visual warnings." |

1. Reproduce: Run python3 app.py. Notice there's no way to see which items are low on stock.
2. Prompt: Ask the agent to add the alert system and a new dashboard route.
3. Verify: Visit /dashboard to see the new low-stock warnings in action.

| 04 | Refactoring | Cleanup | "@processor.py refactor this messy calculation logic into clean, modular functions. Remove the global state variables." |

1. Reproduce: Read processor.py. It's a "spaghetti" monolithic script with global variables that are hard to test.
2. Prompt: Ask the agent to modularize the logic.
3. Verify: Run the refactored code and ensure calculations remain identical, but the code is now testable and clean.

| 05 | Testing | Quality | "Write a complete pytest suite for this Finance API, covering edge cases like divisions by zero and negative currency values." |

1. Reproduce: Run pytest. It will fail because there are no tests or they are incomplete.
2. Prompt: Use the agent to generate a robust test suite for the app.py logic.
3. Verify: Run pytest again and see all tests pass, including the complex edge cases.

| 06 | Migration | Modernization | "Migrate this legacy Flask app to Python 3.12 standards. Replace old formatting with f-strings and add comprehensive type hints." |

1. Reproduce: Run python3 legacy_script.py. Observe the old % string formatting and lack of type safety.
2. Prompt: Ask the agent to migrate the logic into a modern, type-hinted Flask API.
3. Verify: Run the new Flask app and confirm the output matches the legacy script but with a modern architecture.

| 07 | Bug-Fixing | Diagnostics | "Find and fix the 3 bugs in the Inventory API: the ZeroDivision error, the shared order state, and the discount math error." |

1. Reproduce: Run python3 app.py. Submit an empty order (ZeroDivision). Submit two orders in a row (Order Bleeding). Add a loyalty discount and notice the incorrect total ($0.10 instead of 10%).
2. Prompt: Ask the agent to "Find and fix all logic bugs in the 07-bug-fixing/app.py file."
3. Verify: Use curl or Postman to submit orders and ensure the calculations are accurate and orders don't persist across requests.

| 08 | Security | Hardening | "@app.py identify and patch the SQL Injection and XSS vulnerabilities. Use parameterized queries and proper output escaping." |

1. Reproduce: In the search box, type ' OR '1'='1 (SQLi) or <script>alert('XSS')</script>.
2. Prompt: Use the agent to audit app.py for vulnerabilities and apply security best practices.
3. Verify: Attempt the same attacks; they should now be safely handled/escaped.

| 09 | Performance | Optimization | "@data_proc.py optimize the sales reporting logic. Replace the O(n^2) nested sorting loop with an efficient algorithm." |

1. Reproduce: Run the analytics report and notice the significant lag (3+ seconds) caused by the inefficient sorting loop in data_proc.py.
2. Prompt: Ask the agent to optimize the data processing logic for better time complexity.
3. Verify: Run the report again; it should be near-instantaneous.

| 10 | Deployment | Ops | "Generate a multi-stage Dockerfile and a GitHub Actions workflow that runs tests and builds the image on every push." |

1. Reproduce: Inspect the existing Dockerfile. It's massive and unoptimized.
2. Prompt: Ask the agent to create a slim, multi-stage Docker build and a CI/CD pipeline.
3. Verify: Build the image locally (docker build . -t workshop-app) and check the image size reduction.

• Context Loading: In Gemini CLI, always start by loading your file context with @.
• Iteration: Agentic coding is a conversation. If the first result isn't perfect, just follow up: "That looks great, but could you move the button to the header?"
• Shell Tasks: Don't exit the agent to run tests. In the CLI, use !pytest. In Antigravity, the agent can run commands for you!

Happy Coding! 🚀