🎨 Scenario open-sourced an entire creative production toolkit that lets Claude Code, Codex and Cursor operate like a small AI-powered 3D studio.

Instead of using a different AI tool for every step, you can give your agent a creative brief and let specialized skills coordinate concept art, 3D modeling, animation, materials, lighting and game-engine integration.

The project is called Scenario Agent Skills.

And it's much more than a collection of image-generation prompts.

What can your AI studio do?

▪️ Art direction: Generate character concepts, creatures, moodboards, storyboards and consistent visual references.

▪️ 3D modeling: Turn sketches and reference images into textured meshes using models such as Meshy, Rodin and SPAR3D.

▪️ Animation: Auto-rig characters, extract motion from videos, retarget animations and prepare clips for game engines.

▪️ Materials and texturing: Generate seamless PBR textures, apply materials, adjust shaders and maintain consistent visual styles.

▪️ Environment creation: Build skyboxes, props, walkable rooms and larger 3D environments.

▪️ Asset optimization: Perform retopology, UV unwrapping, texture baking, LOD preparation and export.

▪️ Lighting and rendering: Configure Blender Cycles, EEVEE, Maya Arnold or real-time game-engine lighting.

But the architecture is the interesting part.

Scenario combines two different types of agent skills.

1. AI generation through Scenario MCP

The agent connects to Scenario's model catalog and uses specialized skills to:

▪️ Select the appropriate AI model for each task.
▪️ Inspect supported parameters and reference inputs.
▪️ Estimate credit costs before generation.
▪️ Generate images, 3D assets, videos or audio.
▪️ Preview the results and identify problems.
▪️ Apply targeted corrections instead of regenerating everything.
The system also includes identity libraries for maintaining consistent characters and visual styles across multiple assets.
2. Expert tools for professional software
Separate skill families let agents operate applications installed on your computer:
▪️ Blender: Modeling, sculpting, rigging, animation, Geometry Nodes, UVs, rendering and exports.
▪️ Maya: Character animation, skinning, grooming, simulations and production pipelines.

▪️ ZBrush: Digital sculpting and mesh preparation.

▪️ Unreal Engine: World building, materials, Niagara VFX, performance optimization and automation.

▪️ Unity and Godot: Gameplay systems, 3D environments, animations, shaders, UI and asset pipelines.

These skills use Python scripts, application APIs, command-line tools and local MCP bridges.

Here's a practical workflow.

Imagine you're building a stylized game.

You ask:

"Design a medieval blacksmith character, generate a textured 3D model, prepare it for animation, create an idle cycle and export it for Godot."

Your agent can coordinate:

Character concept generation.

Image-to-3D reconstruction.

Mesh cleanup and retopology.

PBR material preparation.

Rigging and animation.

GLB export and engine integration.
You review the results and request corrections between stages.
Quality control is built into the process.
Scenario provides skills for comparing models, inspecting generated assets, checking consistency and iterating against predefined quality criteria.
The expert toolkits also include application-specific validation procedures.
Installation:

npx skills add scenario-labs/skills

Choose the relevant skills, then connect Scenario MCP if you want cloud generation.

The expert tools for Blender, Maya and game engines can operate separately using local software.

What does it cost?

The skills themselves are open source under MIT.

However:

▪️ Scenario cloud generation consumes credits.
▪️ API access requires a paid Scenario plan.
▪️ Professional applications may require separate licenses.
▪️ AI agent subscriptions and compute costs are additional.

And this isn't a guaranteed one-prompt production pipeline. Complex assets still require review, debugging and artistic decisions.

Why this matters:

AI 3D generation is improving rapidly, but producing an attractive mesh is only one part of creating a usable asset.

Game development also requires clean topology, textures, animation, optimization and engine compatibility.

Scenario is trying to connect those steps into a coherent agent-driven workflow.
The future of AI-assisted creative production might look less like one giant generator and more like a coordinated team of specialized tools.
💻 https://github.com/scenario-labs/skills
🌐 https://scenario.com/

#AIAgents #3DArt #GameDev #OpenSource