cc-blender-skill
Thirty chained skills drive Blender through MCP: modeling, materials, lighting, camera, rendering, animation and glTF export
- 159 ★ on the repository
- Last commit
Broad access
Why this level
- The skills execute generated Python inside Blender and can change or overwrite scene files
- The agent reaches Blender through an MCP server, and some scenarios download assets from the network
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/cc-blender-skillDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The skills execute generated Python inside Blender and can change or overwrite scene files
- The agent reaches Blender through an MCP server, and some scenarios download assets from the network
This is third-party code. Review the repository files before installing.
What it does
cc-blender-skill is a set of thirty skills that turn a sentence like "build a sword and render it with three-point lighting" into Python for Blender. The `text-to-blender` orchestrator recognizes the task and loads the right sub-skills: modeling, materials, lighting, cameras, rendering, animation, export, `wireframe-to-3d`, UV unwrapping and result validation. The code runs in Blender through `mcp__blender__execute_blender_code`. Two server routes are supported: the official Blender Lab MCP (needs Blender 5.1 or newer) and ahujasid MCP (Blender 4.0 and newer). The chosen server stays fixed for the whole task, and tasks that download or generate assets go through ahujasid from the start. The set targets Codex and Claude Code; the content is shared while installation and MCP setup depend on the agent. Version 1.4.0 added dual MCP support, Unreal collision and Multires cage export. `wireframe-to-3d` has its own skill that builds parametric 3D models for glTF and GLB from orthographic drawings.
Who it is for. 3D artists and technically minded designers who already work in Blender and want to build scenes, materials and renders with text requests to an agent.
Good fit when
- You need to build an object from a description and get a render with lighting and camera set up right away
- You have an orthographic drawing of glasses or a part and need a parametric model for glTF
- A model is being prepared for Unreal and needs collisions and Multires cage export
- You want a fixed order: modeling first, then materials, unwrapping and result validation
Not a fit when
- You need a ready MCP server for Blender: the skills only guide the agent and the server is installed separately
- You need Blender below 4.0: the official route needs Blender 5.1 or newer and the ahujasid route 4.0
- You need fully unattended generation: the author lists separately what does not work yet and results need a human look
Example request
Make a teapot model in Blender and render it with three-point lightingLimitations
You need Blender, a configured Blender MCP server and a compatible agent. Installation and MCP setup depend on the host, and you will follow the guide in `plugin/skills/text-to-blender/references/agent-hosts.md`. The Python code runs in your Blender and can change the scene or project files.
The author splits features into working and not-yet-working ones and reports validation gaps in the v1.4.0 integration report. The symlink install into `~/.agents/skills` refuses to replace existing skills with the same names. Some scenarios on Blender 4.x are not live-tested.
How to disable. Remove the skill symlinks from `~/.agents/skills` (names match the folders in `plugin/skills/`) or delete the installed plugin, then remove the Blender MCP entry from the agent configuration.
FAQ
Which MCP server does Blender need?
The official Blender Lab MCP with Blender 5.1 or newer, or ahujasid MCP with Blender 4.0 or newer. Tasks that download assets go through ahujasid.
What does wireframe-to-3d do?
Turns orthographic 2D drawings, for example eyeglasses in two views, into a parametric 3D model optimized for glTF and GLB export.
How do I install the set in Codex?
Clone the repository and link folders from `plugin/skills/` into `~/.agents/skills`, or into `.agents/skills` in a project. The README install command refuses to replace existing skills.
Does the set work in Claude Code?
Yes, the author lists Claude Code as a supported host. Instructions are in the `agent-hosts.md` guide inside `text-to-blender`.
How is this different from the Blender MCP server itself?
The MCP server gives the agent access to Blender while the skill set tells it in what order and with what techniques to build models, materials, lighting and renders.
Discussion
No questions or reviews yet.
Related
SkillsBlender Skills
blender-skills
Blender skills for Cursor, Claude Code, Kiro and Codex: work goes through the BlenderMCP add-on that has to be installed in Blender
A skill collection guides AI agents through Blender modeling, materials, animation and asset checks using supplied Python scripts.
MCP server and addon for Blender 5.2: 276 tools for modeling, materials, animation and rendering with mesh quality checks
Blender MCP
MCP for Blender
An MCP server and Blender addon: the agent creates and edits objects, materials and scenes, pulls assets and runs Python in Blender
An enhanced Blender MCP build: work with a live Blender and with saved .blend files without a UI, runtime bpy API lookup and the BlenderBench test suite
A small Node.js Blender MCP: a stdio server sends Python code to Blender over TCP, reads the scene and pulls Poly Haven assets
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| cc-blender-skill (this card) | no | 159 | ||
| Blender Skills | no | 2,557 | a simpler set without an orchestrator or two MCP options | |
| Blender MCP | no | not required | 34,097 | the MCP server itself, not a skill set on top of it |
README in short
The repository holds a portable skill pack for Codex, Claude Code and compatible hosts. The plugin lives in `plugin/`, research material in `knowledge/` and `docs/`. The author separately documents verified behavior and known gaps of version 1.4.0.
README badge
Original SKILL.md
# Wireframe-to-3D Conversion Skill **Name**: `wireframe-to-3d` **Purpose**: Convert orthographic 2D wireframe images (technical drawings) to parametric 3D models in Blender, optimized for glTF/GLB export. **Use Cases**: - CAD-style eyeglasses from front/side orthographic views - Technical wireframes of mechanical parts - Industrial design sketches with dimensional profiles - Character feature accessories (goggles, helmets, etc.) **Status**: Ready for integration --- ## Quick Start ### For Users ``` You: "Convert my wireframe glasses drawing to a 3D Blender model" Skill response: 1. Analyze the wireframe image(s) you provide 2. Extract contours and fit Bezier curves 3. Create parametric curves in Blender 4. Convert to mesh with proper topology 5. Apply materials and export GLB → Returns: model.glb + log of optimization steps ``` ### For Developers The skill orchestrates three stages: **Stage 1: Image Analysis** (`wireframe_analyzer.py`) - Input: PNG wireframe image - Output: JSON with Bezier control points - Runtime: < 1s for typical 800×600 wireframe **Stage 2: Blender Curve Creation** (Blender MCP via Python) - Input: Analyzer JSON output - Output: Blender `.blend` scene with curves - Runtime: < 2s