by DY-2026
Local-first OS for AI-assisted game design: turn AI-agent sessions into decisions, evidence, experiments, proposals, and durable project memory.
# Add to your Claude Code skills
git clone https://github.com/DY-2026/GameDesignOSGameDesignOS is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by DY-2026. Local-first OS for AI-assisted game design: turn AI-agent sessions into decisions, evidence, experiments, proposals, and durable project memory. It has 181 GitHub stars.
GameDesignOS's catalog security scan is still queued. You can run an instant dependency and prompt-injection check now with the "Scan for vulnerabilities" button above.
Clone the repository with "git clone https://github.com/DY-2026/GameDesignOS" and add it to your Claude Code skills directory (see the Installation section above).
GameDesignOS is primarily written in Python. It is open-source under DY-2026 on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh GameDesignOS against similar tools.
No comments yet. Be the first to share your thoughts!
Unlocks once the catalog security scan passes (runs nightly).
The deep catalog scan for this skill is still queued. Run an instant dependency check now instead.
License: skill documents and tooling are released under the MIT License. The Paranoia name, logos, visual identity, and project branding are not licensed as trademarks.
GameDesignOS is the public-facing system name. Paranoia is the author identity and brand signature; the current installable modules remain packaged as Markdown skills.
AI can draft faster than a team can decide. In game design, that often creates more fragments: chat logs, screenshots, one-off prompts, competitor notes, GDD drafts, experiment ideas, and half-remembered decisions.
GameDesignOS adds the missing operating layer. It turns agent output into reviewable project assets with provenance, contracts, gates, and rollback, so a designer can move from idea to evidence to experiment to decision without rebuilding context every session.
You still make the calls. The system supplies specialist workflows, shared handoffs, local state, and the discipline to stop at Human Gates.
| Category | Count | What It Gives You |
|---|---|---|
| Specialist skills | 7 | Concept architecture, experience analysis, ED optimization, proposal writing, workflow evolution, book translation, and source curation |
| Contract schemas | 17 | Stable handoffs for decisions, assumptions, evidence, experiments, learning, gates, workflows, issues, player promises, and project assets |
| v1 workspace sections | 9 | A durable private project space for Decisions, Assumptions, Evidence, Experiments, Design Assets, Workflows, Learning, Exports, and runtime state |
| Workflow guides | 5 | Idea-to-validation, media-to-diagnosis, weekly ED experiment, evidence-to-proposal, and decision-to-information paths |
| Host adapters | 4 | Codex, Claude Code, OpenAI-compatible agents, and local harness integration notes |
| Public proof cases | 2 | Evidence-linked game-analysis and experience-density examples with explicit source boundaries |
| Runtime | 1 | Deterministic local CLI for routing, workspace creation, validation, health checks, graphs, gates, and review-safe packs |
v1.0.0 turns the local gamedesignos runtime into a formal Project-Ready design operating system. The CLI is deterministic and local-first: it creates v1 workspaces, manages Decisions, Assumptions, Evidence, Experiments, Gates, Workflows, and Learning, exports decision graphs, scans project health, and builds review-safe packs without calling a model.
| Layer | Purpose | Entry |
|---|---|---|
| Skill Kernel | Seven bounded specialist workflows | Current Skills |
| Contract Layer | Stable handoffs, schemas, and routing boundaries | contracts/ |
| Project Workspace | Durable decision, assumption, evidence, experiment, workflow, and learning assets | runtime/workspace-template-v1/ |
| Runtime Interface | Executable local commands plus host-agent integration boundaries | runtime/ / gamedesignos/ |
Start a private workspace:
python -m pip install -e .
gamedesignos "I want to make a lighthouse tactics game"
The natural-language entry recommends the right skill. For project-shaped requests, it prepares the workspace, first Decision, first Assumption, three-minute validation Experiment, VOI Gate, and workflow in one pass. Then run the small test and use the printed gamedesignos evidence add ... command to record the observation. Existing skill folders remain independently installable; the runtime remains compatible with v0.8/v0.9 workspaces.
Read the v1.0 development plan, CLI guide, command reference, and product roadmap.
Decision-first research prompt:
Use $paranoia-ai-system-evolver to audit this research or AI workflow with a Decision Object, current default action, decision boundary, EVPI/EVSI, signal-to-action map, the smallest high-VOI probe, and a stop rule.
GameDesignOS is a local-first operating system for AI-assisted game design. It turns AI-agent sessions into durable design assets: decisions, assumptions, evidence, experiments, proposals, workflows, and learning records.
Its public base consists of a Skill Kernel, Contract Layer, Project Workspace, and executable Runtime Interface for concept validation, gameplay diagnosis, proposal writing, and workflow evolution.
The skills provide bounded expert behavior. Contracts make outputs interoperable. The workspace preserves project context. The runtime layer gives a host agent or local CLI deterministic commands for reading, routing, writing, validating, packing, and stopping at Human Gates.
It is not a scattered skill list or a prompt dump. It is closer to a compact operating system for serious game design work, with contracts that let skills hand work to each other instead of producing isolated prose:
media evidence -> evidence index -> issue cards -> ED handoff -> weekly experiments
one-line idea -> player-promise contract -> validation plan -> later media diagnosis
concept/evidence/production notes -> decision-ready proposal -> pitch or milestone gate
workflow change -> WOOP Task Card -> VOI/OODA probe -> eval -> Human Gate -> rollback
books and sources -> structured knowledge assets -> better references for future work
SKILL.md, references/, templates/, evals, Human Gates, and rollback paths.needs_review; real projects stay in your own environment.Use $game-experience-analyzer to diagnose this PV or gameplay recording into sample boundary, timestamped evidence, Hook/Loop/Link/Surprise diagnosis, issue cards, and validation recommendations.
Use $game-concept-architect to turn this one-line game idea into concept seed extraction, design nucleus options, player promise contract, core loop, scope gate, and prototype validation plan.
Use $game-design-proposal-writer to turn this concept brief, validation plan, evidence notes, and production constraints into a decision-ready commercial proposal, indie dossier, publisher pitch, or vertical-slice document.
Use $paranoia-ai-system-evolver to upgrade this workflow with a WOOP Task Card, VOI, OODA, eval checks, Human Gate, and rollback.
Use $game-experience-density-optimizer to turn this first-session retention, pacing, or experience density problem into an ED diagnosis, CLP/SF/EB/AR/MD-min levers, a weekly A/B plan, instrumentation, dashboard fields, decision rules, and rollback gates.
For the full onboarding path, see Try It in 10 Minutes. Star this repo if you want more public game-analysis examples and portable agent-skill templates; more stars help prioritize better public dem