Bias Dungeon
Pixel-Art Educational Browser Game · Built Apr 2026
A 5-minute pixel-art browser game that teaches players to spot cognitive biases through everyday decision traps. Players move through five dungeon rooms, make choices, reveal the bias behind each situation, and leave with one practical anti-bias move.
The Problem
Most cognitive bias education is academic, passive, or too abstract. People may know terms like anchoring, confirmation bias, or sunk cost, but still miss them in everyday moments: online shopping, social media, gossip, navigation apps, or decisions about time and money.
My Role
Solo builder leading the full product from idea to shipped MVP: concept, PRD, architecture, game loop, UI, implementation, testing, and deployment. I built the project during a live Vibecoding event with Marily Nika, Ph.D on Maven.
What I Built
- Built a complete 5-room educational game that turns cognitive bias theory into interactive decision traps.
- Designed a simple learning loop: everyday scenario, player choice, bias reveal, and practical anti-bias move.
- Shipped a polished browser experience with pixel-art UI, retro sound, animation, recap screen, and Vercel deployment.
Technical Architecture
- Built a client-side React + Vite + TypeScript single-page app with no backend or database dependency.
- Created a typed content and game-state layer for rooms, choices, reveals, avatars, and progression.
- Structured the game around a clear state flow: start, avatar select, room intro, reveal, recap.
- Used Tailwind CSS, Framer Motion, Howler.js, GitHub, and Vercel to ship a responsive static web app.
AI Workflow & Tools
- Claude Projects for product planning, PRD structure, and scope definition.
- Claude Code and Cursor for implementation, debugging, refactoring, and shipping the MVP.
- Perplexity AI for cognitive bias research and scenario grounding.
- Google AI Studio for UI prototyping and creative exploration.
Results
- Shipped a complete playable MVP in 4 hours during a live Vibecoding event.
- Turned an abstract education topic into a hands-on game people can test immediately.
- Published the project as both a live Vercel app and an open-source GitHub repository.
Key Learnings
The biggest learning was speed: with the right AI workflow, the gap between idea and working MVP can shrink dramatically. In four hours, I moved from concept to a playable product, but the product judgment still mattered most: knowing what to simplify, what to ship, and what made the learning experience memorable.