This is a curated collection of twelve classic arcade games (1971-1981) recreated for the browser using nothing but vanilla JavaScript, HTML5 Canvas, and the Web Audio API. Each game lives in its own directory with a strict three-file structure: index.html, style.css, and a single game.js that contains all logic. The code is clearly authored with educational intent and an emphasis on historical authenticity.

The 'problem' it solves is preservation and education: it lets anyone play and study faithful recreations of foundational video games without emulators, ROMs, frameworks, build tools, or external dependencies. Each game aims to reproduce original hardware behavior, scoring systems, and CRT-style visuals rather than being merely 'inspired by' the originals.

The audience is hobbyist developers, retro-gaming enthusiasts, and people learning game programming. Usage is trivial: clone the repository and open any game's index.html directly in a browser — no server, no install step. The single-file game.js design also makes each game a readable, self-contained learning resource.

Each game.js follows a consistent architecture organized into labeled sections: a frozen CONFIG object of tunable constants, math utilities, sprite/shape data, a procedural sound engine, input handling, entity classes, collision detection, rendering, a game state machine, and a fixed 60Hz accumulator-based main loop. A small Python helper script uses PIL to generate a composite screenshot grid from each game's screenshots, parsed chronologically from the README.

No red flags. The code is benign client-side game logic with no network calls, no data collection, and no obfuscation. The repository includes a .claude/skills directory and an IMPLEMENTATION.md, suggesting it was built with AI assistance following a repeatable game-building template — consistent with the highly uniform structure across all games.