game-cog

**Quick judgment** `game-cog` is a strong, specialized skill for end-to-end game content generation, best suited for teams or solo devs who want cohesive worlds (art + design + 3D + UI) from high-level briefs. It builds on the `cellcog` stack, using agent/daemon patterns to handle heavy, long-running asset generation. **Key strengths** - Holistic game-focus: not just sprites, but full GDDs, worldbuilding, UI, and 3D assets tied to a single vision. - Good practical guidance: concrete example prompts and mode guidance (`agent` vs `agent team`) for different workloads. - Engine-aware outputs: GLB models and tilesets with control over poly count/materials, easing integration into common engines. **Risks & limitations** - **Dependency on `cellcog`:** Requires installing and understanding the base `cellcog` skill; failures or config issues there will break `game-cog` workflows. - **Quality variance:** As with any generative pipeline, asset quality, style fidelity, and animation usability may vary and will likely need human review/iteration. - **Latency & cost:** Multi-step generation (concept → reference art → 3D models, or complex GDDs) will be relatively slow and compute-expensive, even with async patterns. - **Content control:** While the domain is comparatively safe, teams still need guidelines to avoid unwanted violent or inappropriate content, especially for younger audiences. **Recommended scenarios** - Indie or small studio teams needing fast pre-production: concept art, GDDs, and placeholder assets to validate ideas. - Tooling for rapid prototyping game jams and pitches, where cohesive but not final assets are acceptable. - Technical designers who already use `cellcog` and want a higher-level, game-specific orchestration layer for multi-asset generation. - Education and experimentation, to teach or explore game design workflows from ideation through basic asset creation.