💡 Summary
A retro-style arcade space shooter game developed using Flutter and AI-assisted techniques.
🎯 Target Audience
🤖 AI Roast: “This game is like a retro blast from the past, but with AI doing the heavy lifting—talk about a power-up!”
The README does not indicate any direct security risks, but the use of external dependencies like Flutter and the AI agent could introduce vulnerabilities. Regularly update dependencies and conduct security audits to mitigate risks.
Space Shooter
A retro-style arcade space shooting game built with Flutter and Flame Engine.
Built with AI: This game was developed using self-evolving-agent's
/evolveskill - an autonomous AI agent that learns and iterates until goals are achieved.
Play Now
Screenshots

Features
Weapon System (5 Types × 4 Levels)
| Type | Color | Special Ability | |------|-------|-----------------| | Standard | Green | Balanced power | | Spread | Yellow-Green | Wide coverage | | Laser | Cyan | Piercing (hits multiple enemies) | | Missile | Orange | Homing (tracks enemies) | | Plasma | Purple | High damage |
Each weapon has 4 upgrade levels increasing bullet count and power.
Power-ups (12 Types)
Basic Power-ups (20% drop rate)
| Item | Color | Effect | |------|-------|--------| | Weapon Upgrade (W) | Orange | Level up current weapon | | Shield (S) | Blue | 8s protection | | Rapid Fire (R) | Purple | 5s double fire rate | | Speed Boost (>) | Yellow | 6s faster movement | | Bomb (B) | Red | Clear all enemies | | Extra Life (+) | Green | +1 life |
Weapon Type Power-ups
| Item | Color | Effect | |------|-------|--------| | Spread (SP) | Yellow-Green | Switch to spread weapon | | Laser (LA) | Cyan | Switch to laser weapon | | Missile (MI) | Deep Orange | Switch to missile weapon | | Plasma (PL) | Violet | Switch to plasma weapon |
Formation Bonus
| Item | Color | Effect | |------|-------|--------| | Mega Bonus (★) | Gold | +500 score + random power-up |
Enemy Types (8 Types)
| Type | Shape | Behavior | Spawn Direction | |------|-------|----------|-----------------| | Basic | Hexagon (Red) | Sine wave movement | Top | | Fast | Triangle (Orange) | High-speed straight | Top/Diagonal | | Tank | Square (Gray) | Slow but 3 HP | Top | | Shooter | Diamond (Purple) | Fires bullets | Top | | Zigzag | Star (Cyan) | Side-to-side movement | Top | | Tracker | Pentagon (Pink) | Follows player | Any direction | | Diver | Arrow (Yellow) | Dive attacks | Top | | Orbiter | Circle (Blue) | Orbital movement | Left/Right |
Spawn Directions: Top, Left, Right, Top-Left, Top-Right
Enemy Formations (5 Types)
- V-Shape: Classic formation
- Line: Horizontal row
- Circle: Circular pattern
- Arrow: Pointed formation
- Wave: Sine wave pattern
Formation Bonus: Destroy all enemies in a formation to get guaranteed bonus drops!
Boss Battles (Enhanced)
- Appears every 5 waves
- 3-Phase difficulty system:
- Phase 1 (>70% HP): Basic patterns
- Phase 2 (40-70% HP): More bullets, faster fire
- Phase 3 (<40% HP): Enraged mode with spiral attacks
- Enraged Mode (<30% HP): Speed boost + aggressive patterns
- Health: 1.5-2x higher than before
- Drops 3 random power-ups on defeat
Controls
| Key | Action | |-----|--------| | ↑↓←→ / WASD | Move | | SPACE | Shoot | | ESC | Pause |
How This Game Was Built
This project demonstrates AI-assisted game development using the self-evolving-agent workflow:
Step 1: Create Domain Skill
First, a Flame Engine skill was created based on the Flame Engine documentation. This skill encapsulates:
- Flame architecture patterns (FlameGame, Components, World/Camera)
- Input handling (keyboard, touch, joystick)
- Collision detection system
- Sprite animation patterns
- Audio integration
- Common game patterns (object pooling, spawners, parallax)
Step 2: Evolve the Game
With the Flame skill in place, the /evolve command was used to iteratively build the game:
/evolve 建立一個太空射擊遊戲 /evolve 加入武器升級和道具掉落系統 /evolve 加入多種敵人類型和 Boss /evolve 部署到 GitHub Pages
The self-evolving-agent autonomously:
- Searched memory for relevant patterns
- Applied Flame Engine best practices from the skill
- Implemented features with proper architecture
- Tested and fixed issues iteratively
- Deployed the final product
Development Time: ~40 minutes (initial) + ~20 minutes (enhancements)
Initial build (~40 minutes):
- Game architecture setup
- Player with 4-level weapon system
- 6 types of power-ups with drop system
- 5 enemy types with unique behaviors
- Boss battle system
- Wave progression
- HUD with lives display
- GitHub Pages deployment
Enhancement update (~20 minutes):
- 5 weapon types with unique properties
- 3 new enemy types with multi-directional spawning
- Enhanced Boss with 3-phase difficulty
- 12 power-up types
- Formation bonus system
Bug fixes (2026-01-13):
- Fixed Bomb power-up not clearing enemies (recursive search for nested components)
- Added visual confirmation flash effect when Bomb triggers
Why This Approach?
| Traditional | AI-Assisted | |-------------|-------------| | Read docs → Write code → Debug | Skill captures expertise → AI applies patterns | | Manual iteration | Autonomous PDCA cycles | | Knowledge in developer's head | Knowledge encoded in reusable skills | | Hours to days | ~40 minutes |
Skills are "packaged judgment" - they tell AI when to use what patterns, reducing decision points and ensuring consistency.
Tech Stack
- Flutter 3.x
- Flame Engine 1.x
- Dart 3.x
- AI Agent: self-evolving-agent
Development
# Install dependencies flutter pub get # Run in development flutter run -d chrome # Build for production flutter build web --release
Deployment
The game is deployed to GitHub Pages using the gh-pages branch.
# Build and deploy flutter build web --release # Copy build/web/* to gh-pages branch
License
MIT License
Credits
- Built with Flame Engine - A minimalist 2D game engine for Flutter
- Developed using self-evolving-agent - An autonomous AI development workflow
Pros
- Engaging retro-style gameplay
- AI-assisted development reduces coding time
- Diverse weapon and enemy types enhance replayability
- Easy deployment via GitHub Pages
Cons
- Limited graphics and sound design typical of retro games
- May not appeal to players preferring modern graphics
- AI dependency could limit customization
- Potential bugs from rapid development cycles
Related Skills
pytorch
S“It's the Swiss Army knife of deep learning, but good luck figuring out which of the 47 installation methods is the one that won't break your system.”
agno
S“It promises to be the Kubernetes for agents, but let's see if developers have the patience to learn yet another orchestration layer.”
nuxt-skills
S“It's essentially a well-organized cheat sheet that turns your AI assistant into a Nuxt framework parrot.”
Disclaimer: This content is sourced from GitHub open source projects for display and rating purposes only.
Copyright belongs to the original author miles990.
