agentleFS
Sign inStart free
artmediastudioartmediastudio/kimi-blender-mcp-plugin/skills/blender-mcp/SKILL.md

blender-mcp

Skill4 starsChanged 4 months agoMITView on GitHub

Control a running Blender instance through the blender-mcp MCP server. Use when the user wants to create, inspect, or modify 3D scenes in Blender: modeling, materials, lighting, cameras, rendering, scene queries, viewport screenshots, running Python (bpy) inside Blender, or importing assets from Poly Haven, Sketchfab, or Hyper3D Rodin. Triggers on requests like "create a 3D scene/object in Blender", "connect Kimi to Blender", "model this in 3D", or any Blender automation task.

  • Installs packages

5 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions5 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 5

    1. oneshot-repo/OneShot★ 48same name
    2. vladmdgolam/agent-skills★ 9same name
    3. Remote-Arc/remote-arc★ 5same name
    4. yaohuangguan/remote-arc★ 5same name
    5. leonardoaraujosantos/my_ai_skills★ 1same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 3 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Verify first: call get_scene_info before anything else.

    line 33
  2. Very specific:

    Look before editing: use get_scene_info / get_object_info to ground every change.

    line 36
  3. Very specific:

    Right tool for the job: prefer dedicated tools for queries and asset imports.

    line 39
  4. Distinctive:

    One client at a time: never connect a second MCP client (e.g.

    line 46
  5. Distinctive:

    Sketchfab: search and download models.

    line 54
  6. Distinctive:

    If an integration tool fails with an auth error, ask the user for the key — do not retry in a loop.

    line 58

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Blender via MCP

Drive Blender through the MCP tools exposed by the blender-mcp server. The server talks to the bundled Blender addon over a local socket (port 9876).

Setup (one-time, on the machine running Blender)

  1. Install Blender 3.0+ and the uv package manager (pip install uv or see https://docs.astral.sh/uv/).
  2. Register the MCP server (stdio transport):
    • Kimi Code CLI: kimi mcp add --transport stdio blender -- uvx blender-mcp
    • Windows fallback if uvx is not found: use command cmd with args ["/c", "uvx", "blender-mcp"], or the absolute path to uvx.
  3. In Blender: Edit → Preferences → Add-ons → Install… → select assets/addon.py from this bundle → enable "Interface: Blender MCP".
  4. Open the 3D viewport sidebar (press N) → "BlenderMCP" tab → start/connect the server. Blender must stay open while you work.

Working rules

  • Verify first: call get_scene_info before anything else. If it errors, the addon is not connected — walk the user through setup step 4. Never fabricate scene state.
  • Look before editing: use get_scene_info / get_object_info to ground every change; take a viewport screenshot after structural changes to confirm the visual result when the user cares about look.
  • Right tool for the job: prefer dedicated tools for queries and asset imports; use execute_blender_code for creation, transforms, modifiers, materials, lighting, cameras, and rendering. Write idempotent bpy code (delete or reuse existing objects by name instead of duplicating them).
  • Safety: execute_blender_code runs arbitrary Python in Blender — always tell the user to save the .blend file before destructive or experimental operations, and confirm before deleting objects.
  • One client at a time: never connect a second MCP client (e.g. Claude and Kimi simultaneously) to the same Blender instance.
  • Iterate in small steps: build scenes incrementally, checking each step, instead of one giant script.

Asset integrations

  • Poly Haven: HDRIs, textures, and models — free, no key needed.
  • Sketchfab: search and download models; gated content may require an account/API token.
  • Hyper3D Rodin: text/image-to-3D generation; requires an API key.

If an integration tool fails with an auth error, ask the user for the key — do not retry in a loop.

Troubleshooting

  • Connection refused / timeout → Blender closed, addon disabled, or the BlenderMCP panel not connected (setup step 4).
  • uvx: command not found → install uv, or use the Windows cmd /c form.
  • Tools appear but do nothing → another MCP client is holding the connection; close it and reconnect from the Blender panel.

For deeper server details see https://github.com/ahujasid/blender-mcp.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free