Donchitos/Claude-Code-Game-Studios

Turn Claude Code into a full game dev studio — 49 AI agents, 72 workflow skills, and a complete coordination system mirroring real studio hierarchy.

View on GitHub ↗Jump to charts ↓Open shareable report

Summary Information

Updated 2 minutes ago
Added to GitGenius on April 23rd, 2026
Created on February 12th, 2026
Open Issues & Pull Requests: 58 (+0)
Number of forks: 3,499
Total Stargazers: 24,488 (+1)
Total Subscribers: 180 (+0)

Repository Insights (GitGenius)

Median issue/PR response: 2.1 days
Mean response time: 6.6 days
90th percentile: 23.5 days
Tracked items: 26

How this project is maintained

Around half of the issues opened in the past year never receive a reply. 100% of open issues come from outside the core team, so the backlog reflects real-world use rather than internal planning. Only 6% of issues opened in the past year have been closed. Three people close 60% of everything that gets resolved.

Charts & Analytics

Fetching additional details & charts...

Issue Activity (beta)

Open issues: 34
New in 7 days: 0
Closed in 7 days: 0
Avg open age: 24 days
Stale 30+ days: 28
Stale 90+ days: 16

Recent activity

Opened in 7 days: 0
Closed in 7 days: 0
Comments in 7 days: 0
Events in 7 days: 0

Top labels

  • enhancement (13)
  • bug (7)

Most active issues this week

Detailed Description

Claude Code Game Studios is a comprehensive framework that transforms a single Claude Code session into a fully structured game development studio with 49 specialized AI agents, 73 workflow skills, and 12 automated hooks. The repository is written primarily in Shell and addresses a fundamental problem in solo AI-assisted game development: the lack of organizational structure that leads to technical debt, inconsistent design decisions, and poor code quality.

The core innovation is organizational hierarchy mirroring real game studios. Rather than relying on a single general-purpose AI assistant, the system provides three tiers of agents including creative and technical directors who guard the overall vision, department leads who own specific domains like programming and art, and specialists who execute hands-on work. This structure ensures that decisions flow through appropriate channels, quality gates are enforced, and cross-domain changes are properly coordinated through a producer agent.

The 73 skills cover the complete game development lifecycle. Onboarding skills like /start guide users based on their current project stage, from no idea through existing work. Design skills include /brainstorm, /design-system, and /propagate-design-change to maintain consistency across the game vision. Development workflows are supported through /create-epics, /create-stories, /dev-story, and /story-done commands that structure work into manageable units. Specialized domains have dedicated skills: /art-bible and /asset-spec for visual assets, /ux-design and /ux-review for interface work, /qa-plan and /regression-suite for quality assurance, and /team-combat, /team-narrative, /team-ui for coordinating multiple agents on specific features.

The system includes 12 automated hooks that enforce quality standards without requiring manual intervention. Hooks validate commits for hardcoded values and proper TODO formatting, check asset naming conventions, detect gaps in documentation when code exists, maintain audit trails of agent activity, and preserve session state across compactions. These hooks run automatically on session lifecycle events and tool usage, providing continuous safety guardrails while failing gracefully if optional validation tools like jq are unavailable.

Engine-specific support is built in for Godot 4, Unity, and Unreal Engine 5, each with dedicated specialist agents and sub-specialists for engine-specific concerns like GDScript and shaders for Godot, DOTS/ECS for Unity, and GAS for Unreal. The framework includes 41 document templates covering game design documents, UX specifications, architecture decision records, sprint plans, and accessibility guidelines.

The collaboration model is explicitly non-autonomous. Agents follow a strict protocol of asking questions before proposing solutions, presenting multiple options with pros and cons, drafting work for review, and requiring user approval before finalizing changes. This keeps the human user in control while providing structured expertise and catching mistakes early through domain-specific review processes.