I vibe coded a mobile game in 8 working days. The decision log was the real product.

Description and use

The author built a mobile game in 8 working days using AI agents and a detailed decision log. The process involved clarifying the goal, brainstorming, spec’ing, ticketing, expert review, and incremental implementation. The 396‑entry decision log became the project core and a guide for future rebuilds.

Problem

Developing a mobile game with AI agents while ensuring quality, reproducibility, and maintainability, without relying on human code authorship for every decision.

Outcome

The workflow produced a fully functional mobile game in 8 working days, with a comprehensive decision log that became the core of the project brief, enabling future rebuilds and tutorials.

Workflow

• Clarify the rough idea and ask the agent to restate the goal, list assumptions and unknowns, and ask questions until nothing is ambiguous.
• Brainstorm several approaches for each open question.
• Spec the result into a spec with acceptance criteria.
• Cut the spec into small tickets that are buildable, testable and mergeable on their own.
• Have specialist agents (architect, security engineer, product manager) review the spec and tickets before any code exists.
• Implement one ticket, run a code review, commit, and move on to the next.

Evidence

The repository holds the whole process, not just the code: the original rough idea, four rewrites of it, the spec, 48 tickets, every prompt in order (workflow.md), and a log of 396 decisions.

Source

The workflow is described in full in CodingAgentOrchestration. In short: Clarify, Brainstorm, Spec, Tickets, Review, Build.

Source

Every non‑obvious choice gets a numbered entry with three parts: the question, the decision, and what it costs.

Source

After 391 decisions I asked the agent to rewrite the idea document with every decision folded into it, and the war stories turned into general rules.

Source